1 Task: Predicting traffic
To goal of this task is to construct a model that predicts traffic (number of cars) crossing Gamle Nyg˚ardsbro at a given time.
1.1 Data
Statens vegvesen has a traffic monitoring station at Gamle Nyg˚ardsbro right next to the Department of Informatics. In this project, we use hourly traffic data from December 2015 to December 2019. We are going to predict three different quantities: The total number of cars crossing Gamle Nyg˚ardsbro (Volum totalt), the number of cars towards city center (Volum til SNTR), and the number of cars towards Danmarksplass (Volum til DNP). The features are year (˚Ar), month (Ma˚ned), day (Dag) and time (Fra_time).
The data file data.csv can be downloaded from MittUiB.
1.2 Preprocessing
The first step in the modelling process is to preprocess your data so that you can input it to a learning algorithm. To this end, you need to extract some features from the date.
Remember feature engineering. Use your domain knowledge about the traffic patterns. What kind of variations there are?
Think whether a feature should have a continuous or categorical value. It may make sense to represent categorical features with more than two possible values using one-hot encoding.
Hint: Visualize data to see what kind of patterns there are.
1.4 Bonus task
This task is not compulsory.
Find additional data and use it make a better model.
