F1 Data Junkie

WRC Portugal, 2018

2018-05-18T23:51:00.001+01:00

Rally report: https://wrcportugal2018.rallydatajunkie.com/

F1 Spain 2018 - Practice / Quali / Race Reports

2018-05-11T09:52:00.001+01:00

F1DataJunkie reports for the F1 GRAN PREMIO DE ESPAÑA EMIRATES 2018 can be found here: https://esp2018gp.f1datajunkie.com/.

WRC Argentina Rally Report

2018-04-28T00:10:00.000+01:00

WRC Argentina Rally Report

F1 Azerbaijan 2018 - Practice / quali / race reports

2018-04-27T22:58:00.000+01:00

F1datajunkie reports - Azerbaijan 2018 - Practice / quali / race

F1 2018 China - Practice / Quali / Race Reports

2018-04-14T12:27:00.000+01:00

F1DataJunkie reports for the FORMULA 1 2018 HEINEKEN CHINESE GRAND PRIX can be found here: https://chn2018gp.f1datajunkie.com/

Bahrain F1 2018 - Practice / Quali / Race Reports

2018-04-07T17:18:00.001+01:00

F1DataJunkie reports for the 2018 Bahrain Formula One Grand Prix can be found here: https://bah2018gp.f1datajunkie.com/

Tinkering with chart variants...

2018-03-28T23:42:00.000+01:00

As I try to tidy up automated chart generation for pushing race reports to Gitbook - https://www.gitbook.com/@f1datajunkie - I've also started tinkering with some new charts, such as this variant on the chart to compare the qualifying performance of drivers in the same team which looks at how the drivers in all the teams compared in the qualifying session of a single race:

I'll try to start pushing some of the code to updated chapters in the Wrangling F1 Data With R book too, time allowing... Another reason to buy into that book now is that at some point I want to make a start on a second edition, with a focus on Tidyverse recipes rather than the rather more old-fashioned R the current book contains. Because LeanPub now charge to create new books (at least until an author generates $10k in sales), the new edition will replace the old one at the same URL. (So if you buy now, you'll get the first edition and the second one - if I can find the time to do it! - for one purchase: the LeanPub model means you get all updates for free...)

Publishing F1 reports on gitbooks....

2018-03-24T11:30:00.001+00:00

I'm going to start trying to publish race reports on gitbooks - first attempt, for Australian F1 2018 qualifying, is at https://f1datajunkie.gitbook.io/aus-2018-gp or via the f1datajunkie.com domain at https://aus2018gp.f1datajunkie.com.

All the standard quali reports are there - progression session utilisation, cut-off time evolution, etc.

Initial Peak at Some WRC Telemetry Data

2018-02-28T14:33:00.000+00:00

A few weeks ago I was tipped off that telemetry feeds for WRC rally stages were available on the WRC Plus site, so I had a peak and now I'm starting to have a play...

For example, here's a fragment from Rally Sweden 2018 SS9, Meek vs Ogier, looking at throttle, brake and speed:

Here's where it seems to be on the map...

Not sure what do now, though.... What sorts of thing would be interesting to look for from a rally sports reporter perspective?

And what sorts of thing do the teams look for?

F1 2017 Malaysia - Race Summary

2017-10-02T11:51:00.000+01:00

F1 2017 Italy Race Review

2017-09-04T11:09:00.001+01:00

Pit Activity:

Hmm, looking at that chart, I think I need to: extend the lap range to the duration of the race; in the stacked bar below, order the stacked bars so the stops are ordered left to right; provide an option to rearrange the driver order to rank drivers according to the preliminary classification.

F1 Italy 2017 Qualifying

2017-09-03T13:09:00.000+01:00

A quick summary of how qualifying progressed following the session restart...

Overall rank progress:

Overall session times:

Session Utilisation:

Q1:

Q2:

Q3:

Python and R Wrappers for the ergast F1 Results Data API

2017-07-14T00:04:00.002+01:00

A week or so ago, Will Vaughan tipped me off to a Python package he's started putting together to wrap the ergast motor racing results database: ergast-python.

He's also making a start on some Wrangling F1 Data Jupyter notebooks that make use of the wrapper: wranglingf1data.

Very nice:-)

It also prompted me to finally split out my own crappy ergast API wrapper R code into an R package of its own: ergastR.

It needs tidying up, and could probably do with being updated to "new R" and a more recent JSON package, but it should work enough to cope with my own Wrangling F1 Data With R code. (If it doesn't, shout!)

The R package also needs work on the following: cacheing, so that we don't keep hitting the ergast API unnecessarily; paged results handling (I fudge this a bit at the moment by explicitly setting a large results limit); and dual handling of ergast API versus ergast database requests. (If a database connection string is passed, use that rather than make a call to the ergast API.)

Any issues with the R package, please feel free to raise them on the Github site, although I can't guarantee how quickly I'll get to them...

F1 Austria Race Charts

2017-07-10T10:22:00.000+01:00

Race history chart:

Track position chart:

F1 Austria Qualifying Report

2017-07-09T00:41:00.001+01:00

Qualifying summary:

Qualifying summary - relative laptimes:

Session utilisation:

Q1 cutoff time evolution:

Q1 equivalent grid:

Q2 cut-off time evolution:

Q2 equivalent grid:

Q3 front-row cut-off time evolution:

Q3 equivalent grid:

F1 2017 Azerbaijan - Looking for Stories...

2017-06-26T00:04:00.001+01:00

It's one thing to watch the race, it's another to look at the charts to try to spot things that may be of immediate interest. So from a quick skim of the charts, which ones jumped out at me?

For a null point starter for 10, the track position chart. From lap 35 or so, we see the race evolve at the front, if nothing else...

In terms of battlemap charts, Bottas has one of the more interesting ones... Can you just feel yourself holding your breath from lap 39 or so...?

Hamilton was casually hassled for the first half of the race (not that interesting, though?) and hassling from lap 30 or so:

So how do Vettel see it? Close fought throughout?

Stroll was battling from the off (but there's a lot of boring safety car stuff in there in the first half of the race?), makes space on MAG behind, and then BOT starts to threaten... Gulp... How much per lap..?

Looking at the charts, you'd maybe credit VAN with a charge?

So.. the race from Bottas' perspective looks like it may have stories to tell... How does it look from the track concordance perspective?

Fun...?!:-)

And what about Ricciardo? How did he go about getting from 19th to the top step of the podium?

Passes on MAG, HUL and STR maybe...? But then he was gifted?

Thinks... it would be really nice to link those two charts together interactively, so that highlighting an element on one chart highlighted the other instances of that driver on the same chart and the other chart?

F1 2017 Azerbaijan Qualifying

2017-06-25T13:21:00.001+01:00

How the session progressed:

How they qualified:

In terms of time:

Going from Q1 to Q2:

KVY failed to improve his time, recording 103.186 compared to 102.927 (0.259s slower)
HUL failed to improve his time, recording 104.267 compared to 103.93 (0.337s slower)
WEH failed to improve his time, recording 104.603 compared to 104.317 (0.286s slower)

Going from Q2 to Q3:

STR failed to improve his time, recording 102.753 compared to 102.284 (0.469s slower)
MAS failed to improve his time, recording 102.798 compared to 102.735 (0.063s slower)
RIC failed to improve his time, recording 103.414 compared to 102.215 (1.199s slower)

How Q1 cut-off time progressed:

As-if time-based grid from Q1 times:

How Q2 cut-off time progressed:

As-if time-based grid from Q2 times:

How Q3 front-row cut-off time progressed:

As-if time-based grid from Q3 times:

F1 2017 Monaco Qualifying

2017-05-27T15:57:00.000+01:00

Qualifying position, slopegraph summary

Session utilisation - when were cars on track during the session?

Cut-off time evolution - Q1:

Session best time, Q1 - grid relative to timing:

Cut-off time evolution - Q2:

Session best time, Q2 - grid relative to timing:

Front-row of grid cut-off time evolution - Q3:

Session best time, Q3 - grid relative to timing:

WRC Rally Portugal 2017

2017-05-20T19:33:00.002+01:00

If you follow WRC, my automated rally report slides for this rally are here: https://psychemedia.github.io/WRC_sketches/wrcTest-slides-Portugal.slides.html#/

F1 2017 Spanish Grand Prix - Race Charts

2017-05-14T18:33:00.001+01:00

(Chart reading based on not watching race, hearing any commentary from it, or having read any race reports - it's just a reading of the charts.)

The race history chart shows a battle up front between HAM and VET but then not much action until the battle for the last few points between two cars for 7th and three cars for ninth until the last few laps. The dip around lap 35 or so suggests a safety car?

The track history chart shows the paucity of cars finishing the race on the lead lap, with multiple cars two laps behind the leaders.

The battlemap for HAM shows him trailing, but keeping pace with, VET for the first 13 laps, drawing awy from BOT at the same time, taking the lead as VET pitted.

At lap 22, with the race history chart showing HAM pitting (I need to add pit events to the track charts...), HAM he made significant ground on VET before VET regained the pace. On lap 44 or so, HAM, regained the lead, then slowly started to open up a small gap before what look like a couple of attempts by VET in the last few laps to make ground before falling back again in the presence of backmarkers.

From VET's point of view, we see him leading from HAM at the start, then chasing down BOT at a rate of knots after pitting.

After passing BOT, he had a quick dance with HAM?, he managed to pull away from HAM until the safety car(?). Did HAM make a pass after the safety car(?)? From lap 16 or so we also see the presence of back markers.

The track concordance chart for HAM shows him keeping pace with the car in front (VET) at the start while pulling away from BOT in third at just under a second a lap and the car in fourth by about 20/15 seconds a lap. In the closing laps, we see how the backmarkers appear to have knocked VET back.

The track concordance chart for VET shows his pace in laps 15-20 chasing down BOT and how he lost out to HAM between lapps 45 to 55 or so.

Having a quick look at GRO's race, we see him battling with SAI in the first part of the race, and KVY in the last third as RIC bears down on him like a steam train. (I can almost hear RIC getting shouting with KVY!) Note to self- need to fix the battlemap colours...two factors are 1 and 2 laps behind and need lap(s) ahead/behind colour style cf track concordance chart.)

KVY's battlemap shows how he was further sandwiched between GRO and MAG in the latter part of the race before MAG fell out of the picture.

From SAI's perspective, we seem him battling directly with MAG in the early part of the race, with GRO biting at his heels, and with WEH for the last half as MAG tailed off then dropped from view.

From WEH's battlemap, he doesn't seem to have had much rest, being largely sandwiched closely from behind, but too far ahead to make for any racing challengin for a better position ahead, for pretty much the whole race.

STR looks to have had an interesting race. Hassled by ERI early on as WEH made a small gap, he then chased down ALO but couldn't get close enough to make a pass. Instead, ERI hunted him down and made the pass. He then chased down ALO for a second time, this time making the pass, and did the same with ERI, who then sat in in his mirrors as KVY pulled away quite a bit further up the road. At lap 42, ERI broke past and charged away.

PER seems to have just been having something of a solo Sunday drive..

On the other hand, ALO had to work for the first three quarters of the race before having a pacey chase down of MAS in the last third.

ALO's track concordance chart is too messy to reveal much?

F1 2017 Spanish Grand Prix - Qualifying Charts

2017-05-14T11:57:00.002+01:00

Qualifying progression:

Session utilisation:

Q1 cut-off time evolution:

Q1 - Grid spacing based on best session time:

Q2 cut-off time evolution:

Q2 - Grid spacing based on best session time:

Q3 front-row time evolution:

Q3 - Grid spacing based on best session time:

For more, see F1DataJunkie F1 2015 Spain Qualifying Report.

To learn how to create this charts and support this project: Wrangling F1 Data With R.

A New Chart Type - Race Concordance Charts

2017-05-08T22:09:00.000+01:00

Although this blog doesn't attract that many comments, one of the repeated ones has been for a wider scope in battlemaps so that they show more than just the lead chasing or chased car.

To a certain extent, the track position charts provide a macroscopic overview of the race as a whole from a track position viewpoint, and race history charts show the relative pace of different vehicles.

But how about a chart that could provide a view of the track, both ahead and behind, within a certain time window, from the perspective of a particular car?

This is where track concordance charts come in. For each lap, these charts show the cars ahead - and the cars behind - a particular target car, giving you an at a glance view, on a lap by lap basis, of the traffic ahead and behind the target car, on track.

Here are a few examples of original sketches from the 2017 Russian F1 Grand Prix:

The circles represent cars ahead (negative acctimedelta, in seconds, to the left) and behind (positive acctimedelta, in seconds, to the right) the target vehicle. The colour identifies the lap the cars are on - pale blue is the same racing lap, dark blue shows cars a lap ahead (that need to be let by under blue flag conditions), and orange (to red) are backmarkers behind. The highlights within a circle marker are used to denote other target cars. The vertical, y-axis identifies the lap the target car is on.

In the top chart, the ivory coloured vertial bars represent the pit window, a time +/- pit loss time away from the target car. Vertical dashed lines identify the lap or laps on which the target car pitted.

One of the things our eyes detect in the charts are the progressions of a particular car, assuming they lie along a perceived line. (Aren't the principles of the Gestalt Theory of Perception wonderful?!)

We can actually make traces of a particular car explicit by using a line to connect them. Note that we also need to break the line where the trace for a particular car skips a lap, so we don't get a "carriage return" line across the canvas:

We can also play with the window size:

To my eye these charts are harder to read than the simpler scatterplot style race concordance charts. In certain respects, they are reminiscent of race history charts, providing an idea (from the gradient) of particular lines) of the pace of other cars nearby on track over several laps, relative to the pace of the target car. Additional annotations, such as pit lines and pit windows, may add useful value to this chart without further complicating the reading of it (indeed, they may make it easier to read).

For more detail about the origins of the race concordance chart, see: Race Track Concordance Charts.

F1 2017 Russia Race Review

2017-04-30T20:21:00.003+01:00

The charts suggest this was not the most exciting of races...

And the battlemaps...

Making a Start on Some BTCC Charts

2017-04-30T11:33:00.002+01:00

I've been meaning to do a Python version of Wrangling F1 Data With R for some time, but rather than using F1 data, I thought I'd see how far I could get with TSL timing data from the UK BTCC Championship (and support races).

Here's a first quick sketch...

F1 2017 Russia - Qualifying Report

2017-04-29T17:34:00.001+01:00

Qualifying report charts here: http://psychemedia.github.io/wranglingf1datawithr/2017/f1_2017_rus_quali.htmlhttp://psychemedia.github.io/wranglingf1datawithr/2017/f1_2017_rus_quali.html