The annual TOTO Hong Bao draw is coming this Fri, with a $12 million prize for the jackpot. Being a true-blue Singaporean, I have bought some numbers too. We all know that this is a game of chance, luck and rather low probability of winning, but wouldn’t it be interesting to see some real data?
There are two parts of this project. The first is to get the data. 3 years of Group 1 jackpot data are available in Singapore Pool’s website. The extraction was automated using Python, and dumped into a sqlite database. Google Maps API was also used to geocode Singapore Pool branch’s addresses so as to place markers in a map.
With the more time-consuming coding over, all the data were finally visualised using Tableau. And the resulting dashboard below is created. Click on the link to view interactively!
1. Is Quick Pick Better?
There are basically two ways of picking numbers: 1) Ask the system to generate some random numbers (Quick Pick). 2) You choose your own numbers. Well, out of 153 Group 1 jackpot winners, an overwhelming majority of 138 won using Quick Pick. Yea, those numbers that you dreamt of rarely win. You are better off letting the system random-pick for you.
Also, most Group 1 draws are from Ordinary and System 7. But of course it doesn’t really explain the odds since a System 12 will have a much higher chance of winning compared to an Ordinary system. Logically, it can just be explained by the fact that Ordinary and System 7 are much cheaper ($1 & $7 respectively) so more people buy them each time. And of course, with so many buying, the chances of one getting a jackpot is higher.
2. Is there a Pattern in Time?
I would say that there are no seasonal fluctuations when aggregating the data to monthly levels, looking at the bar-line chart. Feb-Apr 2014 seems to have a higher consolidation of people winning though.
3. Is there a Pattern in Space?
While further cluster analysis is possible, a plain look appears to indicate that there are no clustering of Group 1 winners in Singapore. So I guess there are no regions with particularly high fengshu.
4. What are the Locations with Highest Winners?
I’m not too sure what Singapore Pools Account Betting Service is (as there was no address given), which has 6 winners, so we will just skip that. For a better perspective, there are about 300 branches selling toto in Singapore. 131 branches have jackpot winners before, and these are shown in the table in the dashboard. Everstream Trading is the 2nd location with highest wins of 4. Anyone buying from there?
5. What are the Most Common Numbers?
TOTO consist of selecting from a list of numbers from 1 to 49. The barchart showsthe frequency of each number was chosen in a Group 1 draw (including additional 7th number). Does not appear to have any noticeable trends here, but numbers 21 (occurred 60 times), 8 (56 times), 19 (56 times), 11 (54 times), 31 (54 times), 6 (53 times) are the top 6 numbers (note I’m just merely computing the frequency of each individual number occurring in the Group 1 draws).
Excluding the additional 7th number (use the dropdown to select in the interactive dashboard); top 6 numbers are 21 (52 times), 11 (49 times), 19 (49 times), 8 (48 times), 5 (47 times), 40 (46 times).
——
And so there it is. A fun mini-project to play around with and I also picked up a few tricks to improve my coding skills. If anyone win the $12 mil tomorrow, please don’t forget me!
For those into the geek part, you can view the scripts in my github repository.
Map Attack! 
[…] year around the same time, I did a Tableau project on the Ang Bao draw. This year, I decided to do something different, i.e., to create a dashboard on […]
LikeLike