- Junior Developer
- BeCode Company (Ghent)
- 11/04/2022 - 22/04/2022
- Predicting width constrictions during hot rolling.
- Dealing with huge amount of real Data from Production.
- Analysing, cleaning and pre-processing all data files.
- Visualizing data.
- Modelling data using Supervised Machine Learning techniques.
- Predicting width constriction and visualizing it.
-
Main_CoilFile.csv with attributs:
( coil, furnace Number, analyse, Hardness_1, Hardness_2, Width, Temperature before finishing mill, Temperature after finishing mill, Thickness, Thickness profile, c, mn, si, nb, p, s, al, ma, b, n, ti, cr, va, mo ). -
More the 57,000 Csv files of each Coil-file with attributs ( Coil Number, Coil Stage, Lengthpoints and Values ).
- We analyzed, cleaned and processed all raw data.
- We visualised, recognized and calculated an error value representing the constriction for each coil .
- We combined this information in a single dataframe
- From this data we build a model and optimed it
- Put the .csv files with width measurements for each coil in the SignalExport folder.
- Put all additional measurements in the CoilData folder.
- Run the coil_scorer notebook wich will create a dataframe for all coils with their respective scores.
- Then run the correct clean_transform data notebook depending on the choice of score to clean and transform the dataframe.
- At last run the ML-model notebook to create the model.
- To install one of the Libraries , write in the terminal :
└── Python
├── Numpy
├── Pandas
├── Matplotlib
├── Seaborn
├── Sklearn
├── Joblib
├── Typing
└── Os
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent