totvs

Data Challenge

Using the given dataset of transactions (notas fiscais eletrônicas) from a restaurant: https://github.com/TOTVS/MDMStatic/blob/master/code-challenge/TOTVS%20Labs%20-%20AI%20Challenge%20-%20Dataset.zip?raw=true

1. Parse and extract the data.
2. Identify a pattern on any set of fields that can help predict how much a customer will spend.
3. Calculate a sales forecast for the next week.

We recommend that you spend not more than 5 hours on the challenge.

Evaluation criteria

1. Ability to read/parse a dataset from a file.
2. Understanding of common business data.
3. Basic knowledge of artificial intelligence.
4. Ability to build a model to identify correlations and patterns between fields.
5. Code design
6. Ability to write clear documentation
7. BONUS: Ability to build a forecast model.

Submission

1. A link to the github repository.
2. Briefly explain the models that you used/built

####################################################################################

The jupyter notebook file contain the parsing, analysis and modelling of the problem.

If interested on running the notebook, it was set on python=2.7 anaconda virtual env. The only Library necessary to install is lightgbm