This repository is meant to help you start that competition :) Feel free to change the code, add new files and fix possible bugs. You don't have to use it! However, we believe it may be useful.
Repository content:
- hackathon - python module with utils function and a simple parser of the DMOP file. This is a good point to start your project.
- notebooks - jupyter notebooks which show how to read data, parse them, create and exemplary dataset and simple model with Keras etc.
- data - a directory for dataset files.
- Main dataset is located at: https://www.kaggle.com/europeanspaceagency/mars-express-power-hackathon
- FORNAX, augmented dataset (more data) is located at: https://www.kaggle.com/fornaxai/dataadventures
Unpack the files so that folders train_set and test_set reside in the data directory of the project.
To use the provided code with the Main dataset file names have to be changed so that the prefix is omitted,
eg. change test_set/context-2015-01-01_2015-07-01--saaf.csv to test_set/saaf.csv.
Then put the sample_power--2015-01-01_2015-07-01.csv into the test_set folder and rename the file as power.csv.
The code is written in python2.7 (however it should also work with >3 versions) and it was tested on Ubuntu 16.04 system. We suggest you to install the anaconda software which is a simple tool for managing multiple python versions, in case if you are worrying of possible dependency conflicts.
When using the anaconda enviroment management you should start from installing some necessary libraries. Here are the steps:
- Setup a new conda enviroment:
conda create --name hackathon python=2.7 - Activate the enviroment:
source activate hackathon - Install necessary libraries:
conda install pandas numpy tqdm matplotlib - For people who want to install tensorflow with GPU support:
- Install keras library with pip:
pip install keras- do not install keras with conda install, it will automatically install tensorflow without GPU support pip install tensorflow-gpu- For those who do not care about GPU support:
- Install keras with conda:
conda install keras
- Run
pip install -e .in the project root folder.
With this command you can import hackathon module from your python console.
Each submission file will be evaluated against the fourth Martian year's actual average electric current data using the Root Mean Square Error (RMSE) measure.
- ϵ: root mean square error
- cij: predicted value for the ith timestep in the fourth Martian year of the jth parameter
- rij: reference value for the ith timestep in the fourth Martian year of the jth parameter
- N: the total number of evaluated measurements i∈[1,N] with N<=16488
- M : the number of parameters j∈[1,M] with M=33
There are two paths:
- Orbit and Events - place the events in the timeline in context of the orbit
- Mission Plans Visualisation - How to represent a set of operations commands to different subsystems, with time evolution ?