MattLowrieDS

My sports analytics and data science account.

[MattLowrieDS @MattLowrieDS]

I was pretty late discovering the Big Data Cup, but I thought I would try this weekend to see if time series classification could give accurate event prediction when using tracking data as time series input. My initial results are telling me "no" 😞

[MattLowrieDS @MattLowrieDS]

Made it to #NESSIS 2023! My first in person sports analytics conference since 2019.

[MattLowrieDS @MattLowrieDS]

@HemmenKees Landon Donovan vs Algeria WC 2010. We won the group on # of goals. I was in a sports bar that didn't expect so many Americans to show up in the middle of the day to watch soccer. When he scored the place erupted so loud you heard it down the street.

[MattLowrieDS @MattLowrieDS]

Still tinkering with my #BigDataBowl entry (post-deadline). I adjusted my metrics and ranking which now puts PHI as the top scoring OL for pass plays in 2021:
https://t.co/bqOxNgB0el

[MattLowrieDS @MattLowrieDS]

@mrphilroth I guess your right. I just eyeballed the ratio of # of sack/nonsack plays and assumed it was similar. The number of nonsack:sack plays has to be 7793:764 before it is the same ratio of total distances:
https://t.co/iPQSPFDPiw

[MattLowrieDS @MattLowrieDS]

Need help wrapping my head around data from #BigDataBowl 2023. For a part of my submission I looked at total distance of pass rushers for sack vs non-sack plays. Non-sack plays PRs traveled 10x further than sack plays.

[MattLowrieDS @MattLowrieDS]

@mrphilroth But the ratio of non-sack plays to sack plays is similar to total distance, so I expected the avg per play to be similar (higher for non-sack plays)
Number of nonsack plays: 8014
Number of sack plays: 543

Maybe it is just basic math that I'm not comprehending 😣

[MattLowrieDS @MattLowrieDS]

This idea was suggested by @AaronQSmith who mentioned that total distance doesn't account for the difference in number of plays for each team:

[MattLowrieDS @MattLowrieDS]

Could possibly just be an error in my code too, which is published here:
https://t.co/PU6bXoCNaS

[MattLowrieDS @MattLowrieDS]

@mixedknuts @StatsBomb Please have it in Denver! June/July is great here. Actually any time of year is great here!

[MattLowrieDS @MattLowrieDS]

And the code:
https://t.co/bqOxNgAsoN

[MattLowrieDS @MattLowrieDS]

My entry for #BigDataBowl 2023. I found it to be a good challenge this year. I tried a few different approaches but this is the only one I could get a result from in time for the deadline:
https://t.co/w1qGzEsB98

[MattLowrieDS @MattLowrieDS]

The vids!

[Hudl Statsbomb @Statsbomb]

All videos from the StatsBomb Conference are now available to watch on YouTube!

13 of our excellent speakers are available to watch now in one handy playlist, each video full of the latest research and insight in football analytics

Get the talks here ⬇️
https://t.co/7QO6eRESUb

[MattLowrieDS @MattLowrieDS]

Bundesliga is stable, similar to La Liga, but had even less discrepancy between top and bottom. Interesting that the bottom 3rd of teams in every league has consistently lost points.

[MattLowrieDS @MattLowrieDS]

With Messi to PSG, I was curious if there was any more room for the top 3rd of Ligue 1 to take more points. The top teams have been strong so long I expected the avg pts of the top 7 to be flat, but they have been steadily taking pts from the bottom 7 over the last 10 seasons.

[MattLowrieDS @MattLowrieDS]

The Premier League had the highest average point gain of mid-table teams.