Sunday, February 18, 2018

Eugene Loughlin - more family details

Earlier this week I discovered my namesake, and great-great uncle Eugene Loughlin who lived in London, and I also discovered he had a son called Eugene who was five years old in the 1911 census. This census show no details on my uncle - I now know that from the England & Wales, Civil Registration Death Index (1837-1915) he died in July, August, or September in 1907 in Woolwich. Based on the ages of his sons Patrick (6) and Eugene (5) in the 1911 census - they were obviously very young when he died. Patrick and Eugene are my first cousins twice removed - I will try to find out more about them.

Eugene Loughlin married a Marie Quin in October, November, or December 1901 in Greenwich, London. It seems odd for a Catholic family at that time that their first son, Patrick, was not born until 1905 - perhaps there were still births or infant deaths between 1901 and 1905. His death in 1907 would have been at the young age of 37 years - did he die in an accident, get cancer, or die of a heart attack - we'll probably never know. 

It's fascinating to look up ancestors and find snippets of their lives on-line. In another hundred years it will be easy for researchers to do this to find out anything about us today. I have no pictures from the early 1900s to see what my great-great uncle looked like. Future generations will have millions to choose from on Facebook and the like!

Thursday, February 15, 2018

Post #2000 #Blogging

On 13th November 2006, I published my First Blog Post. Today, over 11 years later, this is my 2,000th post. Never would I have thought back then that I would write so much and that it would become part of me. I have written about education, family, sport, travel, books, politics, history, technology, data analytics, and yes - some shite too!

Below I've plotted (using Tableau) the number of posts each month since I started blogging:

Click/Tap image to enlarge.
2016 (in red) sticks out as it was the year I challenged myself to write a blog post every day - which I found difficult to do, and will probably not try again for some time. I can also see it took me a couple of years to get going. After my first post of just two posts in 2006, I only made 15 posts in 2007. 

I've always felt that this blog is a personal website - blogs were originally set up as "web logs", a kind of on-line diary. I often wish I had started doing this when I was much younger - I'd love to reach back to my school and college days (which are fast fading as memories). I have regularly received comments on this blog which have slagged me off, criticised me, disagreed with me, supported me, and of course lots of spam. It can be liberating (and scary) to share thoughts and discuss topics of the day on-line - but what the hell. Not that many people read this blog, though posts do show up in Linkedin views quite a bit. When I write about educational matters it usually gets picked up by, which can double or treble views.

Thanks to all those who have read any of my blog posts. I plan to keep going for at least another 11 years!

Wednesday, February 14, 2018

"Eugene Loughlin" - 1901 UK Census

When I was baptised "Eugene" after Pope Pius XII, it was the first time in some generations that the eldest born male O'Loughlin was not named "Joseph" or "Patrick". My Mum told me that my grandfather PJ (Patrick Joseph) was a little perturbed at the name she had chosen, but could not argue with me being named after a Pope. He then said that he was happy because he had an uncle "Eugene", a brother of his father "Joseph", and that it was a family name after all. 

My great-great uncle Eugene Loughlin was born on 25th November 1870. The "O" and apostrophe was commonly left out in official documents at the time. I am told he was a publican in London, and I decided this morning to search for him in the 1901 census in London using Sure enough, I found him straight away, with his sisters Hanna and Mary, living in 58 Mansford Street in Bethnal Green. There are no houses there now - a block of apartments occupies this site opposite Oaklands Secondary School for girls. Eugene is listed in the census as a "Licenced Victualler" and is classified as "Own account" meaning that he was self-employed. No occupation is given for his sisters. Here's an extract from the census form:

Click/Tap image to enlarge.
I have found no trace of this Eugene after the 1901 census either in UK or Ireland - Perhaps he died between 1901 and 1911. I have found a "Eugene Loughlin" in the 1911 census, aged 5. He had a brother Patrick aged 6 - definitely family forenames. They are listed in the household of Michael and Marie Dennehy living in 206 Albert Road, North Woolwich. This house also does not exist any more, it is in an area that was heavily bombed in World War II - and is very close to the location of an unexploded bomb found earlier this week. Michael is listed as their step father - perhaps a clue that that my great-great uncle Eugene did die between 1901 and 1922. Interestingly, their mother Marie is listed as a "Licenced Victualler" - perhaps keeping up the family business! 

Click/Tap image to enlarge.
More research needed!

Monday, February 12, 2018

How To... Create a Random Sample in Excel 2016 #YouTube

It been a while since I last published a "How To..." video on YouTube (five and a half months to be precise). Today, while browsing through a Research in Education book I came across a short piece on random selection of a sample from a larger population using Excel. I decided to try this for myself and thought such a video might be useful. Using Excel's Data Analysis Tookpak, it is very easy to do - the video is just 3 minutes and 22 seconds long. 

Random sampling is an important technique in Statistics. Every researcher must be careful to avoid bias in sampling, otherwise experiments may be compromised and you may end up with meaningless or misleading results. Here's the video...

Friday, February 09, 2018

Professor Neil J. Salkind RIP #statistics

Image source: Amazon.
For over five years now I have been teaching statistics at NCI, and there's no doubt in my mind that the textbook "Statistics for people who (think they) hate statistics" has been a centre piece for me in preparing for classes, using explanations and examples in my notes, setting tutorials, and getting ideas for exam questions. Salkind's writing style is both humorous and informative - for me he has a brilliant, yet easy, way of explaining statistics so that they can be understood by almost everybody. I have often said to students that his book is my favourite textbook, and if I was ever to write a statistics book I would want it to be like his.

I have never met or communicated with Professor Salkind. Yesterday, while looking up details in Google to order the 6th edition of his book, I was shocked to see an obituary notice for him. He died on 18th November last year. I feel so sad - the world has lost a great teacher. I dedicated my Statistics class last evening to his honour.

Rest in Peace.

Here is Professor Salkind speaking briefly about the 5th edition of his well-known textbook:

Thursday, February 08, 2018

Why is data science sexy? via @james_aka_yale

James Le asks why is data science "sexy" in his on-line article: 16 Useful Advice for Aspiring Data Scientists? In the end he says that "sexiness comes down to being effective".  Hmmmmmm?

Jimmy Floyd Hasselbaink.
Not a Data Scientist!
Image source:
The Dutch footballer Jimmy Floyd Hasselbaink once said, in response to Alan Shearer who claimed that scoring a goal was better than sex:

"You can never say a goal is better than sex - all the guys that say that are not having proper sex."

I guess the same could be said for Data Scientists in what the Harvard Business Review calls The Sexiest Job of the 21st Century. Let's not lose the run of ourselves!

Le goes on to collate advice from 16 data scientists who responded to his question: “What advice would you give to someone starting out in data science?”. For students aspiring to become data analysts/scientists, the list makes for interesting reading. Just a selection of some interesting quotes from Le's article:

"It’s very easy to get a Wikipedia-level understanding of, say, machine learning. For actually doing it, though, you really need to know what the right tool is for the right job, and you need to have a good understanding of all the limitations of each tool."

"put effort into understanding how the data is captured, understand exactly how each data field is defined, and understand when data is missing"

"For the person who’s trying to transition like I did, I would say, for one, it’s hard. Be aware that it’s difficult to change industries and you are going to have to work hard at it."

Joke: "a data scientist is someone who knows more stats than a computer programmer and can program better than a statistician"

"learning how to do data science is like learning to ski. You have to do it. You can only listen to so many videos and watch it happen. At the end of the day, you have to get on your damn skis and go down that hill".

Eugene's advice:
  • Ask a question first
  • Answer the question by using statistics, data mining, and visualization to make sense of the data
  • Think before you plot
  • Challenge every number
  • Above all - be passionate about data!

Monday, February 05, 2018

Re-learning Irish #ConasAtáTú?

DCU are currently running a free on-line course in learning basic Irish - it's called Irish 101: An Introduction to Irish Language and Culture. I decided I'd check it out for two reasons: first as an on-line learning experience - it's only three weeks long and I am less likely to drop out. Secondly, even though I hated Irish at school, I decided to brush up 40 years after my last Irish class and see how much I have to re-learn. 

An Fáinne Nua.
Image source: Gael Linn.
So far, not much of the basics is new to me - but is is really basic in the first lesson, "Conas atá tú?", "Dia duit" and all that. In 1972, after completing a year in a Scoil na nÓg in Trabolgan (Co Cork), I could speak Irish fluently and even had the precious Fáinne Nua to show that I could. But since then I have not had much use for Irish, though like a lot of Irish people I have found it handy when abroad and trying to confuse the locals. I have always been opposed to compulsory Irish in our schools - I prefer it to be a language of choice rather than being shoved down our necks.

The course has a lot of spoken English in it - but that's to be expected for first time learners. I will persevere (I'm almost a third of the way through). DCU have done a good job on this and I hope it is a successful programme for them, and that they plan some more for improvers. Good stuff DCU!

Thursday, February 01, 2018

Being Trolled

Recently a viewer of one of my videos added a comment that simply stated "IRA". Foolish idiot that I am, I responded pointing out that the IRA had nothing to do with the video and that it was a pointless comment. This unleashed several responses such as "Your pointless", "IRA bomber, run!!!!!!!!!", "FUCK THE IRA ACTUALLY NO JUST YOU!", "Over 50 lines of places to bomb!", and "IRA WERE GONNA DIE". Needless to say it is an uneasy feeling that someone would take the time to write comments like this (appalling grammar included) on a video that teaches you how to sort data in Excel. I reported the comments to YouTube as "Spam or Abuse" and blocked the user, but several reappeared as comments. I don't know if they were reposted by the troll, or if YouTube "investigated" the comment and decided they were not "Spam or Abuse".

Image source: Karen Gately.

Compared to some people being trolled - my experience is almost nothing. I often thought to myself that if it happened to me I would brush it off or ignore. I can understand how other people can get upset by abusive comments - it makes me want to reach into my computer screen and give the troll a clip on the ear. There are no more comments today, so I hope the troll has stopped - I hate to think that he (yes - I'm guessing that the troll is male) will have moved on to someone else.

For the record: I have never been in the IRA or planted a bomb.

Saturday, January 27, 2018

Edward Browne, Colour Sergeant

Recently in correspondence with a distant relative, I was shown an extract from a family notebook that appeared to be copies of inscriptions from headstones in St Michael's Cemetery, Gorey, Co Wexford. There are several Brownes buried in this cemetery - see full list at The Brownes are my ancestors on my maternal Grandmother's side of the family. In the note (see middle below), the James Browne that died on 19th September 1876 is (I believe) my Great-Great-Great Grandfather. In the next note (at the bottom of the page), James is recorded as having erected a headstone (presumably) in memory of his father Edward Browne (also from Gorey and who of course would be my Great-Great-Great-Great Grandfather) - to my surprise he is recorded as being a "Coloured Sarg (sic) in Her Majesty's Service"!

There is not much military history in my family tree. While we are sure that my Great Grandfather James Byrne (no relation to the Brownes) served in the British Army in the First World War - no record of him can be found. To have an ancestor as a Colour Sergeant opens up the possibility that he may have seen action in the Napoleonic Wars (1803-1815). I have no record of how old Edward was when he died, but his son James was born in 1811 - so he definitely would have been of fighting age. More research needed. Note that I find there is some doubt about this as the St Michael's Cemetery list indicates that the Edward Browne who died on 15th June 1852, was 46 years old at the age of death.

According to Wikipedia, a "Colour Sergeant (CSgt or C/Sgt) is a non-commissioned title in the Royal Marines and infantry regiments of the British Army, ranking above sergeant and below warrant officer class 2". Wikipedia goes on to say that the "rank was introduced into British Army infantry regiments in 1813 during the Napoleonic Wars to reward long-serving sergeants" and that this rank was considered to be a "a prestigious attainment, granted normally to those sergeants who had displayed courage on the field of battle". Go Great-Great-Great-Great Grand Dad!

Thursday, January 25, 2018

Just how young is our @campaignforleo Taoiseach? #DataViz #Analytics

According a new Irish Times/Ipsos MRBI poll out today, our Taoiseach is enjoying excellent approval ratings of 60%. Pat Leahy writes in the Irish Times that Varadkar’s approval jumps to 60% as voters show their satisfaction. Much of this high level of approval is put down to Varadkar's youthfulness and straight-talking. As I have just started two data visualization modules this week at the College, it got me thinking when I saw this headline how could age of the Taoiseach over time visualized, and how does Varadkar compare to the other Taoisigh?

I looked up Wikipdia and one page gives all the information I needed to draw a time series chart in Tableau. With still quite limited skills in Tableau, I managed to create a coloured area chart with year (1922 to 2018) across the X axis, and age along the Y axis. Each shape that you see represents the time each Taoiseach* served and the age they were each year. The raw data consists of simply listing the name of the Taoiseach for every year between 1922 and 2018 in one column, and their age in a second column. One thing that Tableau is not (yet) good at is adding images. So I had to save the Tableau chart as an image, load it into PowerPoint, insert the images (all taken from Wikipedia), and save the slide as an image (I could have done this in a graphics package either). Below is the result:
Click/Tap to enlarge image.
Indeed, Leo Varadkar is our youngest appointed Taoiseach ever - but not by much. W.T Cosgrave was just 42 when he became President of the Executive Council in 1922, and Bertie Ahern was 46 when he was appointed Taoiseach in 1997. You can see that he has a long way to go to reach the longevity of former Taoisigh - Éamon de Valera was President of Executive Council/Taoiseach for 16 years years from 1932 to 1948!

The Tableau interactive worksheet (without images) for above diagram is available in Tableau Public at this link

*The title "Taoiseach" was not used before 1937, the title "President of the Executive Council" was used from 1922 to 1937 instead.

Tuesday, January 23, 2018

"May the force (of big data) be with you" via @ShradhaMukherj7

Dr Shradha Mukherjee provides us with a really interesting article in the DataCamp Official Blog about the "Top 10 Breakthroughs in Big Data Science in 2017". Google searches for "Big Data" now result in a similar number of results as does searches for "Star Wars". The list of breakthroughs, according to Mukherjee, makes for interesting reading - here are her "Top 10":
  1. Are we alone in the universe?
  2. Wonder how the dolphins are doing?
  3. Will this cancer medication really work?
  4. Can we predict the next severe thunderstorm and tornado?
  5. How to increase safety and minimize crimes?
  6. Is this a new species of plant?
  7. Can molecular structure predict smell of the molecule?
  8. Where did the constitution come from?
  9. Which is the best city to celebrate Christmas in U.S.?
  10. What's the risk of a second heart attack?
All of these questions are being answered with data! The list contains some trivial and serious questions, and I'll not reproduce all the details here - readers can go to the link at the top of this post to read the lot. However, two are most interesting for me: #5 and #9. I never believed that analytics could be used to predict crime, yet police in Vancouver are doing just this. They are using data mining to predict where "crimes are likely to occur" - a valuable tool I think you'll agree in the fight against crime. Imagine police shouting warnings to criminals along the lines of "Armed police - we have you surrounded with data"!

Chicago is the best of 100 cities in the United States to celebrate Christmas. Using 29 parameters, such as "affordability, number of Christmas events, shopping deals", a profile of each city can be built up to come out with the highest score (and best city). The cool thing about this is that any data analyst can do this and create a system of rating cities. You could gather/mine/scrape date from a variety of sources and come up with new and different "Top 100" lists. Imagine doing this for: best city to buy a book in, best city to go to the toilet in, best city for free stuff, and best city for creamy pints of Guinness! The list is endless!

Saturday, January 20, 2018

100 years of Coastal Erosion in Co. Wexford

The Irish Ordinance Survey has a really cool tool, called GeoHive, where you can overlay old maps with modern aerial photography. I have written here before about coastal erosion near our house in Co Wexford, and next week some local residents are meeting to see if we can get some erosion protection works done. Ahead of this I used GeoHive to generate the following image:

Click Image to Enlarge.
The overlay on top of the aerial photo is from the 1888-1913 25 inch map, and it illustrates how much of the coastline has been washed away in just 100 years. I have pointed out where the coastline was in 1988-1913 - this is directly opposite our house. To the left of this is the beach today, (running from top to bottom of photo) with the coastline clearly visible as a band of bushes to the left of this. The area to the bottom right, called "Pattern Green", is almost completely gone today - indeed the last house (at the 7 o'clock position under the "G" of "Pattern Green") made the news two years ago after it was washed into the sea. The photo above must be a few years old, as there has been much more of the coast eroded since it was taken.

The good news is that at the rate of erosion illustrated by my image above, my house will not be washed into the sea for another 100 years!  The bad news is that erosion has accelerated over the past few years and affects the value of every house that you can see above. There's no stopping Mother Nature, and we cannot predict what will happen over the next few years - will Mother Nature use Global Warming to pick away at more of Co Wexford's coast?

Sunday, January 14, 2018

Nothing like a Passport to tell you you're getting old!

I recently had to get a new passport, and there's no doubt that one's appearance changes over the years. In 1974 (at the age of 15) I made my first trip abroad to France - the photo for this was taken by Barney McGirr, the Chemist in Carnew. Of course our passports were green back then - should we join the Brits and hanker nostalgically after our old passports? Every ten years a new Eugene shows up - a little older, and a little wiser. The ten year sequences was broken in 2008 when I lost my 2004 passport in France, and had to get a new one. For the 2018 version the photo was taken with my iPhone - it appears on both booklet and credit card style passports that will keep me going for the next 10 years. 






Tuesday, January 09, 2018

Census 2016 Sapmap Area: County Dún Laoghaire-Rathdown #Census

I often use census data in my Statistics classes (they are a great source of non-normal data) - this year I will be able to use the 2016 census results for the first time. One of the most interesting sets of data is from the Small Area Population statistics - you can look up your county or right down to your local area. The county Dún Laoghaire-Rathdown (DLR) has a population of 218,018 - this includes me. It is also one of the wealthiest parts of Ireland (this does not include me!). Data are easily downloadable in Excel or PDF format.

An interesting analysis to make is to examine how people commute to work. DLR is well-served by public transport with trains DART, LUAS, and plenty of bus lanes. The pie chart below (created in Tableau) shows the breakdown of the census figure for "Population aged 5 years and over by means of travel to work" - this represents 94,397 people.

Data source: Census 2016.
Despite some of the best public transport infrastructure in the country, 47,577 (50.4%) of us still travel by car to work. Train, DART or LUAS is a distant second with 14,094 (14.9%) using this to travel to work. Only 7,781 (8.2%) use buses.

Despite the well-documented increase in the number of people cycling to work, only 5,770 (6.1%) do so in DLR - this is less than the number (6,875 or 7.3%) of people who walk to work. For my own mode of transport (motorcycle) I am one of the 805 (0.9%) people who reported in the census that we use a "motorcycle or scooter" to travel to work.

More to follow!

Monday, January 08, 2018

"Look what a Parking Payment Machine can do!" - Eight years later. #CunningPlan

On April 26th 2010 I wrote a post about the introduction of parking meters on Upper Sheriff Street in Dublin's Docklands (which previously had free parking) - Look what a Parking Payment Machine can do! At the time I thundered "Dublin City Council obviously decided that they were having none of this free parking lark and that folks would have to pay for the privilege of parking on Sheriff Street" and "What genius in the Council decided to do this?". I further moaned "Grown up people in the Council made the decision to charge for parking here, but all they've done is empty the street".

Well - almost eight years later, the street is still empty (the red Hyundai at the right in the photo below is mine at 11:30 this morning). I walk this street a lot at lunch time and it always seems to be empty. I wondered previously if a Cost-Benefit Analysis had been done - you don't have to be a genius to figure that this street is generating very little revenue for Dublin City Council - their cunning plan to generate more money does not seem to be working. 

But... perhaps it does make money? 1 euro buys you just 25 minutes of parking time in this empty street. I used up all the change I had which only got me 50 minutes while I dashed into the College to pick up some papers. 

In 2010 I thought this was " just mean and stupid" - I still do.

Friday, January 05, 2018

Computer science will now be a Leaving Cert subject, via @siliconrepublic #STEM

Yesterday, Education Minister Richard Bruton announced that "an important and timely addition to our education system" is being made with the introduction of computer science in second level schools. According to Eva Short in Silicon Republic, students will be taught "computational thinking and analysis; programming languages and how to modify computer programmes; and how to design webpages, digital animations, simulations, games, apps and robotic systems". Good stuff I hear you say - it should have the desired effect of increasing the number of students studying STEM subjects. There will of course be a knock-on effect for third-level colleges, many of whose introductory computer science modules in first year will have to be pitched at a higher level. This is initially being rolled out to just 40 schools this year, with first Leaving Cert exams taking place in 2020, adding a further complication for third-level colleges as some students coming into first year in 2020 will already have studied computer science, while most will not have.

Interestingly, the Silicon Republic article states that the new course aims to "teach students to be creative, adaptive learners and to employ flexible, solution-oriented thinking" - I would argue that subjects such as Maths, science, and many others already do this. According to the Action Plan for Education (2017), the Dept of Education has the ambition to "create the best education and training service in Europe by 2026" - quite an ambitious target in a European Union with richer countries aiming to do the same thing. No harm in aiming high, but sound-bites like this need to be backed up with action and a lot of money. Hopefully it all comes true.

Image source: Rupert Mallin.
But my main thought on all of this is from the educators perspective. What a time to be a young teacher, or for those studying to become teachers! Richard Bruton is one of the (very few) Ministers I respect - I assume he has the teaching unions on side (and inevitably the money needed too). Many inspiring young educators will be champing at the bit to be teaching these computer science classes. A pity it is just 40 schools first, but rolling this out to all schools at once would be risking too much. It is fascinating to me that in my lifetime, the classroom has changed from using inkwell and pen (yes - I did use these at a desk like the one to the right!), to modern classrooms with a computer on every desk. If every one of our teachers (and children) get this opportunity, there is no doubt that Ireland's future is bright. I might not get to share in this (2026 is the year I formally retire!), but I can watch from the sidelines to see how it all goes.

Wednesday, January 03, 2018

Hans Rosling Animation in Tableau #Analytics #DataViz

I was always fascinated by a TED talk given by Professor Hans Rosling called "The best Stats you have ever seen". In particular his animation of a bubble chart comparing fertility rates, life expectancy, and population for every country from 1960 to 2013. This is the relevant bit in the video (2:39 - 5:02):

I've always wanted to recreate his animation, and now thanks to Kirill Eremenko (via Udemy) - I know how to do it in Tableau Software. It turns out to be a very straight-forward task. I have data for 53 years downloaded from The World Bank - essentially I get Tableau to plot 53 bubble charts and animate them in sequence. Tableau allows a lot of cool filtering too - here's a short video I've made summarising my result:

Saturday, December 30, 2017

A Lesson in Probability #lotto

So - there has been another winner in Ireland of the Euromillions jackpot! Congratulations to whoever it is - their lives have just changed (hopefully for the better). I often use the probability of winning the Irish Lotto in statistics class when discussing the topic of probability. Using combinatorial maths you can work out how many combinations of six numbers there are out of the 47 balls in the Lotto draw - there are 10,737,573. In other words you have a better than a one in ten million chance of choosing the six correct numbers and winning the lotto. As I say to my class, if there was a horse in a race at odds of 10,000,000 to 1 would you put a bet on it? Of course, you do increase your chances of winning if you buy more than one line of numbers.

According to the Irish Lotto website, you have a "1 in 29" chance of winning a prize in the Lotto only draw, (the odds of winning are a better "1 in 17" if you add Lotto Plus). A "1 in 29" chance of winning is equivalent to about 3.5% - if you play regularly you can expect to win about 3.5% of the time, very low odds I think you'll agree. As Charles Wheelan puts it in his book "Naked Statistics", buying a lotto ticket is "a stupid thing to do". Here's how the "1 in 29" ( or approx 3.5%) chance of winning the Irish Lotto is made up:

So if you spend €2 buying one line in the Lotto draw, you can expect a return of around just 7 cent - it's a near mathematical certainty that you will lose money. In last Wednesday's draw (27th Dec), there were 25,213 winners in total of the various prizes. Using the "1 in 29" odds, this means that a whopping 721,177 (25,213 x 29) players did not win!

Of course, somebody does win - IT COULD BE YOU!

Wednesday, December 27, 2017

14,000,000 YouTube views, and who's watching on Christmas Day? #Analytics

On Christmas Day this year, there were 4,101 views on my YouTube channel, which also passed the 14,000,000 lifetime views mark on the same day. December also shows the traditional decline from a high of around 12,000 views per day to a around third of this amount over the Christmas holidays leading to the New Year. 

Click image to enlarge.

So, I wondered who was watching on Christmas day? YouTube Analytics tells me that India leads the way with 1,329 (32.4%) views, while the United States is a distant second with 320 (7.8%) views. In the top ten on Christmas Day are countries like Saudi Arabia (138), Turkey (117), Egypt (85), and the United Arab Emirates (55). None of these countries feature in the top ten list for the whole year so far. Countries like the UK, Canada, Australia, and Ireland basically switch off at Christmas. There were 39 views on my channel on Christmas Day - I wonder if any of these were my students?

The most popular video worldwide on Christmas Day was How To... Perform Simple Linear Regression by Hand with 346 views. Even though there is a question on my exam paper next week on regression (that's not a hint, they already should know this!), in Ireland there was just one view which lasted 4 minutes and 51 seconds - less than half the duration of this video. 

It's also interesting to note that YouTube/Google are now providing more visuals for content creators  to analyse our data - these are visible on the left side panel on the chart above. More on this another day.

Thank you to all my viewers over the holiday season. For those of you preparing for exams I hope that the videos will help in your revision.

Thursday, December 21, 2017

The End is Nigh for Hand-written Exams?

In the age of emailing letters to Santa instead of writing physical letters, Laura McInerney of The Guardian asks "why must students write exams with a pen?". Hardly any of us write letters anymore. Indeed, other than signing your name - when is the last time you actually wrote by hand? I note very few students in my classes actually taking notes with pen and paper. This is unlike when I first came to the College in 2002 when I had to wait to move on slides until students had taken down everything. In fact the only time I see students writing for more than a few seconds is in exams. Some exam papers are extremely difficult for me to read as they are hand-written by students not used to writing for two hours at a time. 

Image source: Huffington Post.

I'd hate to see the skill of hand-writing being completely lost. I agree with Catherine Pearson writing in the Huffington Post about "The Benefits Of Writing With Good Old Fashioned Pen And Paper" - she articulates that as handwriting is slow, it "can be particularly useful during goal setting, brainstorming... — all pursuits that require time and deliberation.". Some famous writers, such as Quentin Tarantino, claim to write all their material by hand as it makes them more creative.

So - should we make students write exams by hand in this day and age? For some subjects, such as programming, practical on-line exams are clearly the best. Many students would prefer to use a computer to write their answers (though I often find that responses are on average a lot shorter than hand-written answer). There are technical challenges, but these are being overcome everyday. Even the great Professor Sugata Mitra (of Newcastle University) "imagines an alternative education system with no need for memorisation or teaching to test" and suggests that a "tablet connected to the internet to be brought in to the examination hall" (see his article in The Guardian "Should students be allowed to use the internet in exams?"). My mind is still open on this, but I see myself favouring computers to be permitted in all exams within a few years. After all, it is the computer that most of us are using at work - not pen and paper!

The good news for old-fashioned lovers of pen and paper is that pen sales are still increasing in the digital age. Sales are expected to reach $20.2 billion worldwide by 2019 according to the Chicago Tribune article "How the pen industry hangs on in a digital world". While this will come under pressure from electronic pens and styluses, the pen is still keeping its "mightier than the sword" status!

Sunday, December 17, 2017

Visit to the City of Preston #PNE

Sir Tom Finney Statue, flat cap, PNE jersey,
I must be in Lancashire!
Yesterday I travelled early to Preston in England via a flight to Liverpool to see the local North End club host Sheffield United with my brother Brian. I have visited Preston on several occasions now (all to see PNE), and I'm beginning to know my way around. Preston people are very pleasant, I love the Lancashire accent where the word "the" is abbreviated to a simple "t". They love their football in Preston and 15,202 showed up to see a competitive game that PNE deserved to win by more than the 1-0 final score. The winner was scored by the £12 million rated Jordan Hugill in the second half. He is a very combative striker who's constantly jostling with his markers. He's not afraid to muscle his way to the ball - he'll be hard for PNE to hang on to in January transfer window if one of the big clubs comes knocking with a fat cheque book. Go PNE!

Incidentally, the return flight from Dublin to Liverpool cost €39 (thank you Ryanair), while parking (short-term) for the day in Dublin airport cost €29 (thank you DAA - not!).

Friday, December 15, 2017

Last Day of Semester #WhoMakesUpThisShit

It's Friday of the last week of semester one and I have now finished all classes - this was the 31st semester that I have completed in NCI. The last week of a semester always gives me mixed feelings in that as each day passes I have a last lecture/tutorial with each class in turn. Some I will see in class again in semester two, others not. They say that time flies if you are having fun, I definitely had a lot of fun working with students this semester, as it absolutely flew by. I also notice that the older I get, the faster the time goes!

Sir Ronald Fisher.
Image source: Wikipedia.
I was asked a lot of questions this semester by students during and after classes - most I hope I gave satisfactory answers to. By far the toughest question I was asked was "Who makes up this shit?"! I had just completed a statistics class and the topic was Analysis of Variance (ANOVA). While the class giggled about this, I was a bit taken aback as I had never been asked a question like this before. I sensed it was asked with tongue-in-cheek, but nevertheless it is an interesting question. Why and how do people come up with new ideas, and who was the first to do something?

A one-way ANOVA is a statistical test to determine if there is a significant difference between the means of 3 or more groups, and it was created by the well-known statistician Ronald Fisher in the 1920s - see a profile of him in Wikipedia. If you want to know how to perform a one-way ANOVA test, check out my video below which shows you the technique that I covered in class - which led to the "Who makes up this shit?" question!

Tuesday, December 12, 2017

Blast from the past #AntiNuke

In a tribute to the late Dr Frank Jeal, posted in the News & Events section on the web page of the Zoology Department in Trinity College, there are many tributes and comments from past students and colleagues of Frank. Included are some photos which bring back a lot of memories of Frank, especially the one below also which includes yours truly (and Frank on extreme left). That combat jacket was one of my favourites, I was also very fond of my head of hair! I'm guessing the photo was taken on a Zoology Field Trip probably in 1980 or 1981 - I don't recognise the building or where it was, though Portaferry in Co Down was the destination for most field trips.

I note in the photo that I am wearing an anti-nuke badge on my left lapel - they were quite trendy at the time as the ESB were planning a nuclear power station at Carnsore Point in Co Wexford. I remember once wearing it to a chemistry lab where our lecturer spotted it and asked me why I was wearing it. No doubt at the time I was worried about nuclear fall-out and Carnsore being wiped off the map (or I was trying to be trendy!) - there was a lot of opposition to it at the time. My lecturer pressed me along the lines of "If it was possible to guarantee that the nuclear plant was 100% safe, would I still be opposed to nuclear energy?" A difficult question for me to answer at the time - he was a PhD in chemistry, I a mere 2nd year student. I'm sure I still opposed it - nothing can be 100% safe. Just 5 years later - the Chernobyl disaster happened.

Monday, December 11, 2017

10 Years of Learning and Teaching with YouTube

On Tuesday 11th December 2007, I created my YouTube channel and uploaded my first video. Given that YouTube was only founded in 2005, and taken over by Google in 2006, I was indeed an early adopter of this medium. 13,875,105 views and 28,603 subscribers later, I am still at it!

My first video was "How To... Convert PowerPoint to iPod Movie". At the time I did not have a SmartPhone (the iPhone was first released on June 29, 2007) - so the old-fashioned iPod seemed to me at the time to be a great tool for learning. In the years around 2007 I mostly taught on the then MSc in e-Learning programme. I created the video as part of an exercise to get students to use technology in innovative ways. Nowadays creating a video is very easy, but back then to get it on an iPod I had to do the following:
  • Create a PowerPoint presentation
  • Import each slide as an image into Windows Movie Maker, add narration, and save as a movie
  • Use a (free) Jodix iPod video converter to convert the video into iPod friendly format
  • Add the video to iTunes, and sync with iPod
This video has been viewed just 6,390 times in ten years, and hardly at all in the past few years as technology is now more enhanced. The picture quality is poor, and some of the syncing of audio to each image is not good. So for a little bit of nostalgia (at least for me!), here is my oldest and first ever YouTube video:

Thursday, December 07, 2017

If you don't know data, you're out of the game. via @tableau

Tableau Software have published "2018: The Year Ahead for Business Intelligence" - it is always interesting to check out what respected and leading companies like Tableau think the future might hold. A key theme throughout is how much easier it is going to be to analyse data so that anyone can do it. While the article is very general, it breaks down into the following 10 topics:
  1. Don't Fear AI
  2. Liberal Arts Impact
  3. Promise of NLP
  4. Multi-Cloud Debate
  5. Rise of the CDO
  6. Crowdsourced Governance
  7. Data Insurance
  8. Data Engineer Role
  9. Location IoT
  10. Academics Investment
Go to the article to read and watch videos (which rather annoyingly are not available to embed) for yourself, but for me two topics stand out for attention: #2 "Liberal Arts Impact", and #10 "Academics Investment". 

Liberal Arts Impact
Who'd have thought that data had anything to do with the Arts? Anya A'hearn (what a brilliant name!) of Datablick tells us that the art of storytelling has helped "influence the data analytics industry" and that "organizations are placing a higher value on hiring workers who can use data and insights to affect change and drive transformation through art and persuasion, not only on the analytics itself" - it's all about telling a story with data.

Academic Investment
A little bit closer to home for me is data analysis, not just teaching, has a role in third level institutions. As Dr Michael Galbreth (University of South Carolina) puts it; graduates "need to be comfortable with data". There is a huge demand from students to learn more about data, with most colleges now having some kind of data analytics/science programme. Colleges are responding to this demand, and we have to be on our game to develop, update, and deliver the right programmes. As Anya A'hearn puts it; "If you don't know data, you're out of the game". True.

Sunday, December 03, 2017

40 Years Ago #ccr

Outside CCR Front Door.
I spent yesterday evening in the wonderful company of about 30 lads in their late 50s who all did their Leaving Certificate in 1977 in Cistercian College Roscrea. We were marking the 40th anniversary of finishing secondary school there. While I see some of the lads reasonably regularly, it was brilliant to meet up with two who I have not seen since June 1977 (PB and TO'T).

We were met in the College for a reception where the biggest treat for me was to meet my French teacher John Shanahan. I told him that I had written about him in my blog when I posted about My Introduction to Learning Technology - September, 1972, and again about My Favourite Teachers - thank you Mrs Keating, Mr Hallahan, Mr Ó Riordáin, Mr Shanahan, Mr O'Loughlin, and Dr Jeal #thankateacherWe had a great chat about Voix et Image and how he was a pioneer of technology in education. Si jamais vous arrivez à voir ce message, merci M. Shanahan pour votre inspiration avec la technologie!

So - walking around the College we were given a tour by the House Captains. A lot has changed in 40 years, though the cold rooms and corridors remain the same. We even attended Mass celebrated by Fr Kevin in the school chapel. In the evening we adjourned to the County Arms in Birr for a super meal, a few beers, and of course more chat. Memories flooded back of our time in CCR 40 years ago - lots of shared stories of bunking mass, robbing orchards, fights, hair cuts, the leather, Roscrea girls, listening to the radio on Saturday afternoons with Tosh, tuck shop, rugby, and of course the bread which kept us all alive. 

I spent just 5 out of my 58 years in Roscrea - yet it is a valued connection when 30 lads can celebrate together and spend an evening in each others company as if we were best friends. We all have had different life experiences with (please God) lots more to come. We vowed we would meet up for a 50th anniversary, though I felt a little sad leaving everybody this morning.

I'd like to give a big shout out to current Roscrea student Manus Heenan who as part of his Transition Year studies has created a mini-business making and selling bread using the original recipe from the monastery. He had the clever idea of selling his goods to a captive audience like us during our reception in the College.  I bought a pack of his ready-to-go bread mix. Details of Manus's business can be found at, and he even has a great YouTube "How To" video showing us how to make Roscrea's famous bread.

Thursday, November 30, 2017

12,000 Views Per Day Barrier Broken #WhosWatching @YouTube

A nice personal YouTube barrier for me was broken this week on Tuesday 28th November when the number of daily views on my channel exceeded 12,000 for the first time (12,167 to be precise). The previous high record of 11,944 views occurred on 25th March, 2014. 12,167 views represents about 21 days and 21 hours of watch time.

The United States continues to dominate with 26% of total views, followed by India (18%) which has been a growing "market" for me in the past couple of years. I have broken the views into ten and in the pie chart below you can see the breakdown by country. The Rest of the Word figure is made up of 146 countries which individually account for less than 1% of views. At the very end of the data are countries like Venezuela, Swaziland, and South Sudan with just one view each last Tuesday.

Tuesday, November 28, 2017

Everything you wanted to know about Data Analytics but were afraid to ask - via @NCIRL

Making the decision to go back to education and study any course is a tough one to make. The commitment and hard work, though definitely worth it, can make for a tough road ahead. However, you won't be alone. At the National College of Ireland most of our students are part-time - attending classes in the evenings and weekends. Data Analytics has become a very popular area of study and most of my teaching now is in this area. As part of the College's efforts to provide as much information as possible to students, we have updated our FAQ section on the College website (see This page now includes a short video of me promoting our next Open Day (Thursday 30th November) - I'll be doing an Information Session at 6 o'clock and will hopefully be able to answer any questions that prospective students may have.

Monday, November 27, 2017

Learning Slow Down During Thanksgiving

It is interesting (to me anyway) to look at the effect of the Thanksgiving Holiday (23rd November) in the United States on the number of views of videos on my YouTube Channel. Since the US accounts for about 27% of all views this year, I would have expected some decline due to the holiday period - this is the case.

In the chart below you can see the effect of Thanksgiving with a marked reduction in total number of views (9,580) beginning the day before Thanksgiving, the day itself (8,224 views), and continuing on Black Friday the day after (7,492 views). This drop is almost exclusively due to the United States. My next biggest markets for number of views: India, UK, and Canada, showed no noticeable drop over the same period. 

It comes as no surprise that less Americans view educational videos during this holiday period. I'm certain that this is reflected in many other areas on YouTube and I'd love to be able to see the data for what goes up, down, or stays the same, over this period. Clearly, Americans (not all - there was still 661 views in the USA on Thanksgiving Day!) are taking time out from learning while feasting on turkey and beer. Learners in the rest of the world (based on just my data) are ignoring this holiday.

November is traditionally a good month for number of views in my channel - the figure for 7th Nov (11,875) is the second highest daily number of views ever for me. Also - the run of four consecutive days of over 11,000 views from 13th to 16th Nov is the first time that this has ever happened. 

Thursday, November 23, 2017

Graduation at National College of Ireland #NCIGrad2017

I am gutted to once again miss the graduation ceremony today of so many of my students from last year - unfortunately classes are clashing with the ceremony. After attending almost every year since I started in NCI, this is the second year in a row that I have missed what had become one of my favourite days in the academic year. 

Image source: University of Economics, Prague.

Many congratulations to all NCI students graduating in the Convention Centre - celebrate the day as you deserve to. In particular I'd like to congratulate students on the following programmes:
  • Higher Diploma in Data Analytics
  • B.A. in Management of Technology in Business
  • B.Sc. in Technology Management
  • Certificate in Business Analysis
  • B.Sc. in Computing
  • B.Sc. in Business Information Systems
  • Higher Certificate in Computing
  • M.Sc. in Data Analytics
It was a pleasure working with you all and I wish you all the best in your future wherever it leads to.

Wednesday, November 22, 2017

What is the point of learning statistics? via @CharlesWheelan

Image source: Amazon.
I am currently really enjoying reading Charles Wheelan's excellent book "Naked Statistics: Stripping the Dread from the Data", and one of his first questions is "What is the point of learning statistics?". It's actually a question I find hard to answer myself - I usually mutter stuff about being better able to understand data and to prove a hypothesis is true/untrue. But Wheelan comes up with excellent reasons why we should study statistics:
  • To summarize huge quantities of data
  • To make better decisions
  • To answer important social questions
  • To recognise patterns that can refine how we do everything from selling diapers to catching criminals
  • To catch cheaters and professional criminals
  • To evaluate effectiveness of policies, programs, drugs, medical procedures, and other innovations
  • And to spot the scoundrels who use these very same powerful tools for nefarious ends
While the first two points above seem obvious, the remainder do not immediately jump to the mind of someone considering studying statistics. 

Studying statistics involves learning a lot of tests - a lot. Figuring out how to set a hypothesis doesn't come easy, and can take a long time to get used to it (I even still struggle sometimes when confronted with a new situation, and I have been teaching this stuff for five years!). But just like everything else, the more you practice, the better you get. Probably for me the biggest thing about statistics is that it gives you value - once you have done the tests, you can analyse the results to make sense of them. Tools such as Excel, R, and SPSS are making statistics easy to do, but you still need to understand and interpret the results.

To finish this post with a quote attributed to H.G. Wells (1866-1946):

Statistical thinking will one day be 
as necessary for efficient citizenship 
as the ability to read and write

Monday, November 20, 2017

Bargain On-line Courses from @udemy #BlackFriday

Last year I purchased two Udemy courses on creating data visualisations in Tableau - both were just $10 as part of the Black Friday Sale. This year Udemy are offering a wide range (over 55,000) of courses once again for just $10 - a no-brainer bargain for learners. For example, one of the courses I am interested in is Building Interactive Graphs with ggplot2 and Shiny, another is Statistics for Data Analysis Using R. Thinking of my own students, there are coursers on Python, R, Data Analysis,  and Statistics - a lot for new learners. Students should certainly consider adding to their portfolio of leaning for just $10 a go.

Image source: Pregnancy and Baby.
Why, may you ask, is a classroom-based Lecturer recommending on-line courses to his students? I already teach Statistics and Data Visualisation modules - it is worrying for students that I am signing up for the above courses? 

The answer is that I am always interested in how others teach in subjects similar to what I do (and of course I am interested in bargains for students as well as myself!). I can learn much from watching others, get ideas for my own class from them, see different examples of data and analysis not covered by the wide range of textbooks that I use, and also get ideas for exam questions. I also almost always find that no matter how often I read a book or watch a video, there is something different that I did not know before which I can add to my knowledge.

These courses can be expensive for students when not on a Black Friday Sale - prices can be up to $145. A huge amount of what you can learn from these courses can be gleaned from a myriad of websites and YouTube videos. But it is a very convenient, and cheap, way to save yourself a lot of searching by signing up for a course with everything in one place. 

I am not associated with Udemy and am not involved in promoting their courses. I just think this is a bargain not to be missed!

Tuesday, November 14, 2017

Six Trends in Data Science

Adam Shapley, writing in Silicon Republic, tells us If you want to be a data scientist, you need to know about these 6 trends. The trends he lists are as follows:
  1. All industries are open, but you should try to specialise
  2. Balance robust academic achievements with on-the-job learning
  3. Data analytics experience is essential, machine learning helps
  4. The GDPR is increasing data governance demand
  5. Make sure you have a solid business intelligence foundation
  6. Keep your technical skills up to date
While I am mostly involved in the education of Data Analysts, this is still an interesting list. I was particularly interested to find that "half of those working in data science have a PhD, whereas less than 2pc of people in the US over 25 years old have a doctorate" (in point 2 above). While a PhD is not a "must-have" for all data science roles, potential employers are sure to take notice. 

Shapley also recommends that Data Scientist learn and maintain news skills regularly. Data science is a complex area, and scientists will need to "demonstrate the most relevant skills and experience to this industry".

Calliostoma zizyphinum (L.).
Image source: UK Natural History Museum.
A PhD can take a long time to achieve - typically 3-5 years. Mine took 4 years and involved a lot of data analysis on shelf shape variation in the painted topshell (Calliostoma zizyphinum) - it takes time to carry out research, analyse results, and write it up. My own thoughts are that a good Masters would also be very valuable in a Data Scientist role - much of course will depend on the level of academic experience sought by employers.

According to Glassdoor, Data Scientist is #1 in the list of Top 50 Best Jobs in America. It rates very high and pays wells ($110,00), today there are 4,184 job openings in the US. A great job like this would make it worthwhile to consider a PhD - even though it could take a long time to achieve, it would be worth it in the long run.

Monday, November 13, 2017

Does anybody trust anybody anymore?

Today I got a letter in the post from my credit card company demanding that under some Terrorist and Money Laundering Act that I provide them with ID and proof of address by return of post. A complete stranger to me will open the letter and enter my details into a computer. Bingo - I am not a terrorist or a money launderer! Yesterday I had to fill out forms to be Garda vetted (yet again) so that I can sing in the church choir.

I was intrigued a few months ago at a presentation about Bitcoin when the presenter told us that the whole idea of Bitcoin and digital currencies is that you trust no one from the beginning. Trust is defined by as: 

"reliance on the integrity, strength, ability, surety, etc., of a person or thing; confidence"

Guilty until proven innocent
Many people are opposed to the idea of being vetted like this: "I never committed a crime in my life" is true of the overwhelming majority of people. It saddens me that elderly people going to church and who volunteer for activities like taking up the collection at Mass now have to be Garda vetted. It saddens me even more that a few perverts have made this necessary. According to the Central Statistics Office, recorded crime incidents classified as "Sexual Offences" were 2,348 offences in 2016, an increase of 8.6% on the previous year. The population of Ireland in 2016 was recorded as 4,757,976. Using the data above, sex offenders make up 0.049% of the population in 2016. It's like trying to find a needle in a haystack using Garda vetting. Yet it has to be done, and is especially important where children are concerned.

The Shoe Bomber
An idiot called Richard Reid, the world's dumbest ever terrorist, tried to blow up a plane with a bomb in his shoe in December 2001. Now we all have to remove our shoes going through security at airports because we cannot be trusted not to do the same. According to the Worldbank, there were 3.696 billion air passengers in 2016, the vast majority of these have to take off their shoes (that over 7.3 billion shoes!). Has anyone been caught with a bomb in their shoe since 2001? Despite this ritual humiliation we all go through at airports, would you get on a plane that passengers were not checked though security?

People have been cheating in exams since forever. But it is just a tiny minority who take the chance to do this. Hence all colleges have strict rules about exams - no phones, (new) no smart watches, no notes, and no water bottles. Every time a student sits down to an exam, he/she has to undergo the ritual instructions from Invigilators. Many of us also make students submit assignments like essays and projects through plagiarism detection systems such as Turnitin. While we can argue that it is part of the learning process to do this, in the end it is about preventing cheating. Why should an honest trustworthy student have to do this?

Trust no one!

Friday, November 10, 2017

Naked Statistics!

Image source: A Little Stats.
Now that the post title has got your interest - read on...! 

I was recently reading in The Economist a review of a book by Charles Wheelan, called "Naked Statistics: Stripping the Dread From the Data". It's quite an old book now (published in 2013), but I've just bought it from Amazon. The piece that made up my mind for me to buy the book was when the reviewer (not named) wrote: 

The reader learns why insurance for low-cost items is worthless and why playing the lottery is a quick way to become poor. More seriously, the book explains the basic statistical approaches used in a 2011 study showing a link between a child’s brain size and autism. And it teems with interesting statistical facts, such as that there may have been an extra 1,000 deaths in the three months after September 11th 2001 because more people chose to drive rather than fly.

The last point is an interesting one and a quick visit to Wikipedia for data on road deaths in the US reveals (using Excel) interesting trends. While roads deaths have fluctuated enormously since 1970, you can see that there is in fact an increase from 42,196 deaths in 2001, to 43,005 deaths in 2002 - an increase of 809 deaths. The trend had been increasing over the previous three years anyway, but there is evidence that road deaths did increase after the 9/11 attacks.  The chart below is really easy to create in Excel (the slowest bit was adding the red label), and I find it is fun to be able to quickly visualise data like this. While road deaths were not caused by the 9/11 attacks (most people of the 43,005 who died in 2002 would have died anyway), it is an interesting thought that in addition to the 2,996 who did die as a direct result of the 9/11 attacks, perhaps the figure should be 3,805 (2,996 + 809)?
Data Source: Wikipedia.