Hands-on for Data Science via Open Education

Ramya N

Published in

Geek Culture

5 min readAug 24, 2021

If you are looking to get into Data Analysis and Data Science

Brief intro: Participation and collaboration with Omdena and Omdena India Chapter

1. Open education — what | by who | how

i. What: It’s a collaborative learning & education approach, works remotely. Helps to upskill and gain hands on experience to build digital, software solution to solve real-world (real-life) challenges through Data Science (AI) tools and technologies.

ii. By Who: Omdena — It’s a collaborative platform and accommodates the above approach.

iii. How: Through this platform you will get to meet collaborators from different places and networking opportunity.

iv. Tools: Python3, Pandas, Plotly Express & Graph_Objects, Dash, Jupyter Notebook

Pandas: Data pre-processing library
Plotly: Interactive Graphing library
Dash: Library to build web app for EDAs that provides interactive features
Jupyter Notebook: Web based editor for Data Science, EDAs and open-source projects

v. If you are looking to get-into AI, upskill or gain hands on experience, I would recommend to get registered with Omdena. I have also provided all the links to visit the Dash web app I built and deployed in ‘Resources’ section at the end, in a sequence. They also have local chapters from different countries.

2. Project I worked on:

Leverage AI to analyze social and economic impacts due to covid19 in India.

We had sub-tasks and active collaborators worked on delivering each part of the project.

Workflow of the project in terms of subtasks.

Exploratory Data Analysis: While exploring and analysing the data, the steps I performed were:

Gathering and collecting data from open-source and authentic resources.
Exploring the required datasets and understanding their fields/columns.
Removing the unnecessary fields and converting the date field to datetime type as required.
Filling the missing values as per the requirements.
Making the fields name consistent if they are not or renaming them as required (for ex: typos).

3. Objectives:

Analyze social and economic impacts due to covid19 in India and to build prediction and forecasting models. The datasets collected from Kaggle, GitHub, Our World in Data, Covid19India and Ministry of External Affairs India.

You can read reference articles in which I have explained about Data Analysis and EDA in the ‘Resources’ section if you are beginner (both these articles are published in Geek Culture publications).

4. What I was involved in:

I participated as a task leader for Data Collection and Pre-processing,
Eventually moved onto EDAs, worked on Descriptive Data Analysis to analyze social and economic impacts.
Built and deployed Dash web app in Heroku.
Moderated the final presentation and presented the demo of the Dash web app.

You can go through the EDA notebooks from my GitHub repo, provided at the end if you wish to visit.

Screenshots of Dash Web App:

Vaccines statistics for first & second doses and covid19 vaccines supplied by India to different countries.

Following code snippets are the examples of data visualization with its relevant output:

Line graph to display total confirmed cases state-wise

Code snippet I have written for the below graph that provides info on total confirmed (delta_confirmed) cases in 2020 — month-wise and state-wise in the Jupyter notebook.

Plotly Express graph to display total confirmed cases in 2020 — month-wise and state-wise.

2. Map graph using Plotly Express and Dash libraries in .py module

To get the map graph for the above dataset, it needs few syntax changes for the Dash library based on specific scenario. Below code snippet is the example of Dash code I did to display the insight through map graph.

2a: Scatter Mapbox type of map graph using Plotly Express and followed by its graph using Dash library to display in web app.

State-wise map graph for total confirmed (delta_confirmed) cases in 2020 — month-wise and state-wise. This graph displayed in web app using **Dash** library.

2b: PIE chart using Python, Plotly and Dash libraries to represent ratio of fully vaccinated people state-wise in India with population more than 20000000.

Pre-processing steps using **Python** and to display in **Dash** **web app**.

**PIE chart** using **Plotly Graph_Objects**.

Similarly I have done the line graph for trend of daily vaccinations in India overall.

5. What I learnt and upskilled:

During this project, learnt about Plotly to make interactive data visualization and graphs along with Dash library to build a web app for EDA notebooks along with hosting it in platform such as Heroku.

It’s a great experience to work on these tools and an opportunity of collaboration and networking.

6. Further scope for this:

User Interface and User Experience can be enhanced for this Dash web app,
Drop-down features can be added for getting drill-down levels of info, ex: state-wise social impact factors can be displayed for both 2020 and 2021 (till date by using real-time data) and month-wise.

7. Resources:

Dash Web app: https://dash-app-eda.herokuapp.com — this app I have built using VS Code and deployed on Heroku. (PS: Reload the browser if it says error as dynos in the Heroku may require to be restarted!)

Video of this application: https://www.youtube.com/watch?v=K9Dv4Bn0MSA

Datasets: Covid19india, MEA and Our world in data

Further Reading: Dash and Plotly

Omdena: https://omdena.com/ — if you are interested to participate you can get more info at this link.

Data Analysis: https://medium.com/geekculture/data-analysis-da-with-python-c570264edeab

Exploratory Data Analysis with Database: https://medium.com/geekculture/eda-with-sql-mysql-4ac1ea1d977b

My GitHub repo: https://github.com/rnedesigns/exploratory_data_analysis

Thank you