I've been making some static plots using GIS to determine what's going on with the trakus_index value as it provides the key to using the GPS latitude/longitude values for deeper analysis.
\nPlotting the final GPS point of the horse who was designated to have finished first, in every case at every track the final location was well past the finish line (see attached image--races are run counter-clockwise in US). I guesstimated where the finish lines were located based on other web sources. Conversely, I checked the final location of the last place horse, only in a small number of races was that location very near to the finish line. In other words, all of the horses have location readings past the finish line and there's no uniformity in distance past finish line.
\nI also plotted start of a particular race ( 1/1/2019, Race 2, 6F on dirt) and found two horses with spurious GPS readings for the first five trakus increments before emitting data that seemed reasonable given the replay video I examined. Another discussion indicated such anomalies should be expected due to the occasional device malfunction.
\nThe question now is what additional information could make the Trakus coordinates analytically useful.
\nbare minimum: latitude/longitude coordinates for the finish lines for all of the courses at the three different locations.
better: some way to correlate race splits with specific trakus_index values. I'm aware that the official Equibase race times may not be equivalent to Trakus increments, but again, we need some sort of context as these coordinates come without timestamps.
Thanks,
\nBrian
I've been making some static plots using GIS to determine what's going on with the trakus_index value as it provides the key to using the GPS latitude/longitude values for deeper analysis.
\nPlotting the final GPS point of the horse who was designated to have finished first, in every case at every track the final location was well past the finish line (see attached image--races are run counter-clockwise in US). I guesstimated where the finish lines were located based on other web sources. Conversely, I checked the final location of the last place horse, only in a small number of races was that location very near to the finish line. In other words, all of the horses have location readings past the finish line and there's no uniformity in distance past finish line.
\nI also plotted start of a particular race ( 1/1/2019, Race 2, 6F on dirt) and found two horses with spurious GPS readings for the first five trakus increments before emitting data that seemed reasonable given the replay video I examined. Another discussion indicated such anomalies should be expected due to the occasional device malfunction.
\nThe question now is what additional information could make the Trakus coordinates analytically useful.
\nbare minimum: latitude/longitude coordinates for the finish lines for all of the courses at the three different locations.
better: some way to correlate race splits with specific trakus_index values. I'm aware that the official Equibase race times may not be equivalent to Trakus increments, but again, we need some sort of context as these coordinates come without timestamps.
Thanks,
\nBrian
What type of GIS are you using? I used time series smoothing from ESRI, and it looks like it worked correctly.
", "rawMarkdown": "What type of GIS are you using? I used time series smoothing from ESRI, and it looks like it worked correctly.", "votes": null } ], "comments": [ { "id": 1990741, "author_name": "thomasclinger", "author_url": "", "post_date": "10/16/2022 18:07:25", "content": "What type of GIS are you using? I used time series smoothing from ESRI, and it looks like it worked correctly.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1955305": "I've been making some static plots using GIS to determine what's going on with the trakus_index value as it provides the key to using the GPS latitude/longitude values for deeper analysis.\n\nPlotting the **final GPS point** of the horse who was designated to have finished first, in every case at every track the final location was well past the finish line (see attached image--races are run counter-clockwise in US). I guesstimated where the finish lines were located based on other web sources. Conversely, I checked the final location of the last place horse, only in a small number of races was that location very near to the finish line. In other words, all of the horses have location readings past the finish line and there's no uniformity in distance past finish line.\n\nI also plotted start of a particular race ( 1/1/2019, Race 2, 6F on dirt) and found two horses with spurious GPS readings for the first five trakus increments before emitting data that seemed reasonable given the replay video I examined. Another discussion indicated such anomalies should be expected due to the occasional device malfunction.\n\nThe question now is what additional information could make the Trakus coordinates analytically useful.\n\n1. bare minimum: latitude/longitude coordinates for the finish lines for all of the courses at the three different locations. \n\n2. better: some way to correlate race splits with specific trakus_index values. I'm aware that the official Equibase race times may not be equivalent to Trakus increments, but again, we need some sort of context as these coordinates come without timestamps.\n\n\n\n\nThanks,\nBrian", "1990741": "What type of GIS are you using? I used time series smoothing from ESRI, and it looks like it worked correctly." }, "source": "meta" }