The main purpose of this project is to predict how many people that will come to work at Telenor Fornebu, for up to approximatly one week into the future. The number of transactions made in the canteens, as well as parking information for the time period October 2016 to February 2019, were provided. It was assumed that the number of people at work equals the number that eats in the canteens. The data set did not contain information about the canteen transactions for all the dates between the start date and end date. Based on the correlation between the parking data and the canteen transactions we were able to deduce how many people were at Telenor when canteen data were missing.
This project assumes you are using Anaconda, with Python version 3.7.3. Use the package manager pip to install requirements.
pip install -r requirements.txtSome may experience problems installing fbprophet, if so install with the following command.
conda install -c conda-forge fbprophetTo get weather data from the Frost API you need to register a profile. The id and secret key from that profile needs to be saved in the config.ini file in the projects root folder. File format example below.
[frost_client]
client_id = x
client_secret = xIf you have the required data files it's possible to run our main jupyter file and play around with it. To update all_models.html after editing the main jupyter file, run the following command.
jupyter nbconvert --to notebook --inplace --execute models/all_models.ipynb && jupyter nbconvert models/all_models.ipynb --TemplateExporter.exclude_input=True --no-prompt --TagRemovePreprocessor.remove_cell_tags='{"remove_markdown"}'