Build a Horse Racing Prediction Model
Budget: $250 – $750 AUD
I am a horse racing enthusiast, and curious about which factors give horses a better chance of winning a race over others.
I would like to establish the probability of a horse winning a particular race using the data I have access to. I understand from others who use horse racing data that a Bayesian approach usually works best, given the nature of the events can change over time.
I have raw csv data going back to 2016, a sample of which is attached. The raw data contains approx 1 million rows and over 150 columns. Each row is unique to an individual horse in a particular race. The data is a mix of numerical and non-numerical data, which can be tested against previous results. An outline of the project is attached with a more detailed explanation of the various columns available on request.
Deliverables
1) The overall objective is to find a probability of the FIN column being WON, 2nd, or 3rd within the context of a particular race. For example, there may be 10 horses in that particular race so we will need to calculate the probability against the direct competition, with the probabilities adding up to 100%.
2) I would like to be able to run any future races myself by running the code in the relevant platform, and also be able to add new data to retrain the model. My knowledge of R and Python are currently very basic but I am prepared to learn enough to run the model, as appropriate.
3) I would like documentation relating to any key findings throughout the project, and reasons for why that particular piece of data was used.
I would like to establish the probability of a horse winning a particular race using the data I have access to. I understand from others who use horse racing data that a Bayesian approach usually works best, given the nature of the events can change over time.
I have raw csv data going back to 2016, a sample of which is attached. The raw data contains approx 1 million rows and over 150 columns. Each row is unique to an individual horse in a particular race. The data is a mix of numerical and non-numerical data, which can be tested against previous results. An outline of the project is attached with a more detailed explanation of the various columns available on request.
Deliverables
1) The overall objective is to find a probability of the FIN column being WON, 2nd, or 3rd within the context of a particular race. For example, there may be 10 horses in that particular race so we will need to calculate the probability against the direct competition, with the probabilities adding up to 100%.
2) I would like to be able to run any future races myself by running the code in the relevant platform, and also be able to add new data to retrain the model. My knowledge of R and Python are currently very basic but I am prepared to learn enough to run the model, as appropriate.
3) I would like documentation relating to any key findings throughout the project, and reasons for why that particular piece of data was used.