Update project to latest libraries and add devcontainer

.devcontainer/devcontainer.json

@@ -0,0 +1,37 @@
+{
+  "name": "data-science",
+  "image": "mcr.microsoft.com/devcontainers/universal:2-linux",
+  "hostRequirements": {
+    "memory": "32gb"
+  },
+  "forwardPorts": [
+    6006,
+    8888
+  ],
+  "customizations": {
+    "codespaces": {
+      "openFiles": [
+        "README.md",
+        "Titanic.ipynb"
+      ]
+    },
+    "vscode": {
+      "settings": {
+        "editor.minimap.enabled" : false,
+        "python.defaultInterpreterPath": "/opt/conda/envs/golden_scenario_env",
+        "workbench.editorLargeFileConfirmation": 1024
+      },
+      "extensions": [
+        "ms-python.python",
+        "ms-toolsai.jupyter"
+      ]
+    }
+  },
+  "remoteEnv": {
+    "ENABLE_ORYX_BUILD": "false"
+  },
+  // There is not a feature for LaTeX in VS Code yet, so we need to install it manually
+  "onCreateCommand": "sudo apt-get update && sudo apt install -y texlive-xetex pandoc && conda env create -f environment.yml",
+  "updateContentCommand": "conda init"
+  // "postCreateCommand": "/opt/conda/envs/golden_scenario_env/bin/jupyter lab --notebook-dir=/workspaces/data-science --ip='0.0.0.0' --port=8888 --no-browser"
+}

.gitignore

@@ -1,2 +1,6 @@
 .DS_Store
-
+.idea/
+.ipynb_checkpoints/
+log/
+.venv/
+venv/

.vscode/extensions.json

@@ -1,6 +1,6 @@
 {
     "recommendations": [
-        "ms-toolsai.jupyter",
-        "ms-python.python"
+        "ms-python.python",
+        "ms-toolsai.jupyter"
     ]
 }

README.md

@@ -8,15 +8,17 @@ This notebook uses machine learning algorithms to get the best accuracy of predi
 - Set the channel priority to strict to avoid issues with the environment creation taking forever.
   - `conda config --set channel_priority strict`
 - Run the following commands (in either the terminal or an Anaconda Prompt):
-  - `conda env create -f golden_scenario_env.yml`
+  - `conda env create -f environment.yml`
   - `conda activate golden_scenario_env`
-  - `conda install python=3.7`
 - In VS Code, open the [Titanic.ipynb](Titanic.ipynb) file and connect to the golden_scenario_env kernel
 
-You need to setup the environment as an `ipykernel` to use it from the Jupyter notebook. To do it run inside of the conda activated environment:
-
-`python -m ipykernel install --user --name golden_scenario_env --display-name "Golden Scenario Env"`
-
 Also if you want to support PDF export from jupyter you need to setup LaTeX:
 
 `sudo apt-get install texlive-xetex texlive-fonts-recommended texlive-plain-generic`
+
+## Dev Containers
+
+You can also run the notebooks inside dev containers:
+
+* [![Open in Visual Studio Code](https://img.shields.io/static/v1?label=&message=Open%20in%20Visual%20Studio%20Code&color=blue&logo=visualstudiocode&style=flat)](https://vscode.dev/redirect?url=vscode://ms-vscode-remote.remote-containers/cloneInVolume?url=https://github.com/claudiaregio/data-science)
+* [![Open in Github Codespaces](https://img.shields.io/static/v1?label=&message=Open%20in%20Github%20Codespaces&color=2f362d&logo=github)](https://codespaces.new/claudiaregio/data-science?quickstart=1&hide_repo_select=true)

environment.yml

@@ -0,0 +1,18 @@
+name: golden_scenario_env
+channels:
+  - conda-forge
+  - defaults
+dependencies:
+  - ipywidgets==8.1.0
+  - matplotlib==3.7.2
+  - notebook==7.0.3
+  - numpy==1.23.5
+  - pandas==2.0.3
+  - pydotplus==2.0.2
+  - scikit-learn==1.3.0
+  - tensorflow==2.12.1
+  - ydata-profiling==4.5.1
+  # pin typeguard to 2.13.3 to avoid a runtime issue with ydata-profiling
+  - typeguard==2.13.3
+  # auto register conda environments as ipython kernels in jupyter
+  - nb_conda_kernels==2.3.1

requirements.txt

@@ -0,0 +1,10 @@
+ipywidgets==8.1.0
+matplotlib==3.7.2
+notebook==7.0.3
+numpy==1.23.5
+pandas==2.0.3
+pydotplus==2.0.2
+scikit-learn==1.3.0
+# it is version 2.13.0 to be compatible with mac too
+tensorflow==2.13.0
+ydata-profiling==4.5.1