3 Python Packages Made For Data Science Teams
Mito is a spreadsheet interface for Python. Many data science teams work with stakeholders who are more used to using Excel. Mito allows users to call a spreadsheet into their Jupyter environment — each edit they make in the spreadsheet will generate the equivalent Python in the code cell below
Here is a demo video:
Mito is not only powerful for Excel users to contribute to a data science team, it also a great tool for Python users. Mito is a faster way to generate code than typing it by hand.
In Mito, you can:
- Import data frames or local spreadsheets
- Create visualizations
- Generate pivot tables
- Filter and sort datasets
- Edit specific cell values
- Merge datasets together
- Export data frames as Excel files
- Automate data cleaning and data analysis processes
- And more!
To install Mito, use these commands:
python -m pip install mitoinstaller
python -m mitoinstaller install
Then open Jupyter Lab and run these commands to call the Mitosheet.
Here is the Mito Team Page, where you can get installation/integration help and custom development, all for free.
Here are the full install instructions.
Plotly is the best Python package for interactive visualizations. You can create UI components to your visualizations where users can change formatting, zoom in on specific pieces of data and more. The goal of many data science teams is to create usable data apps for end users. Visualizations is a key part of this, which Plotly excel’s at.
Here is a demo video:
They also allow users to generate AI and ML focused visualizations. This differentiates Plotly from other graphing packages.
Here is the Plotly Graphing Package documenation.
Here is a link to Dash, Plotly’s enterprise tool.
On the topic of building data apps for end users, Streamlit is an amazing end to end solution for this.
Streamlit allows teams to turn their data scripts into fully functional apps that end users (technical and non-technical) can engage with. Streamlit fully understands that data analysis needs to be presentable and understandable, and they provide an awesome platform to do it.
Here is the Data Professor giving a guide to using Streamlit:
Here is the Streamlit website:
Streamlit can be installed with these commands:
pip install streamlit
I hope you find these packages helpful. I think more Python packages overall need to consider team use cases in their development. I come across many Python packages in my work that would be more useful if they things like sharing features, low-code access, and better support forums.
The packages in this blog are all working hard to make data science for teams an easier process.
In future blog, I hope to explore the different challenges that data science teams face and the work that is being done through different open source and enterprise tools to address them.