COVID19_Vaccine_Hesitancy_Levels_Analysis
Link to Google Slides presentation
The purpose of this project is to analyze and predict the level of Vaccine Hesitancy in USA in late 2021, a year after the outbreak of the COVID pandemic. The datasets which we will be using for statistical calculations and other technology libraries will help identify and display trends, and predict the outcomes accordingly.
-
What counties will rank hesitant versus strongly hesitant to vaccination based on local
public mask mandates,'stay-at-home orders', 'gathering bans' and the prevalence of certainhealth conditions? -
What counties will rank hesitant versus strongly hesitant based on
CVAC levelsmeasuring a county's ability to handle a COVID outbreak? -
What counties will rank hesitant versus strongly hesitant based on its
SVI levelsmeasuring a county's social vulnerability to disaster? -
What counties will rank hesitant versus strongly hesitant based on the local COVID transimission level?
https://data.cdc.gov/Vaccinations/Vaccines-gov-COVID-19-vaccinating-provider-locatio/5jp2-pgaw
https://data.cdc.gov/Vaccinations/Fully-Vaccinated-Adults/jm79-dz78
https://data.cdc.gov/Vaccinations/COVID-19-Vaccinations-in-the-United-States-County/8xkx-amqh
| Software & Libraries | |
|---|---|
| Data Retrieval, Cleaning, and Analysis: | Python, API |
| Database Management/Storage: | PostgreSQL |
| Predictive Analysis: | ML |
| Data Visualization: | Tableau |
-
Data Cleaning and Analysis: Python technology with Pandas library will be used to clean the data and perform an exploratory analysis. Further Trending analysis will be completed using NumPy.
-
Database Storage: PostgreSQL is the database we intend to use, and we will integrate pgAdmin to display the data.
-
Machine Learning: SciKitLearn is the ML library we'll be using to create a regressor. We will split the data into training and testing datasets using a 75-25 split setup. Afterwards, we will use several regression models to validate our machine learning as listed below.
Regression Tree: is similar to decision tree classifier, it runs like a tree with branches to sub-divide all records in training sets and recursively partition until a simple model can fit them. It is suitable for non-linear regression, when interactions are complicated.Linear Regression: Linear Regression is the supervised Machine Learning model in which the model finds the best fit linear line between the independent and dependent variable.Support Vector Regression (SVR): This model improves the flexibility from the linear regression to give an acceptable areas of errors and find a hyperplane to fit the data.
-
Dashboard: Tableau will be used for the final report and integrated with TABLEAU for a fully functioning and interactive dashboard. Finally, the finished data visualizations will be hosted on Google Slides.
Our team members will present a provisional machine learning model that stands in for the final machine learning model and accomplishes the following:
- Takes in raw data from the provisional database
- Outputs label(s) for input data
- Manage repositories rendered by team members
Our team members handling Data Management will present a provisional database that stands in for the final database and accomplishes the following:
- Images of code
- Sample data that mimics the expected final database structure/schema
- Provisional draft of machine learning module with code showing connection to the provisional database
The Regression Tree model we built could predict the vaccine hesitant rate on a county level based on community health conditions, orders, SVI, and supply. It had the best overall predictive performance. It could offer suggestions for the government when campaigning and rolling out vaccines.
Insights: The government's orders and health guidance sent important messages to the public and help the people raise awareness towards the virus. As we see places with or without gathering bans rated significantly differently on their hesitancy. Also, the analysis highlights the social vulnerability index and health conditions as key points to consider when rolling out vaccine. Vulnerable people with fundmental diseases should be cared the most during the pandemic, but the analysis shows they are the most reluctant to vaccination and suffering from high or substantial transmission. The government though priotized seniors and vulnerable ones for vaccination, the analysis suggests more campaigns, guidances, or instructions should be done.
For the further research, we will recommend:
Wider the data scope to include more counties. Sentimental analysis on social media posts, as it also plays a key role in vaccine hesitancy. Scratch and collect news about vaccine and conduct keyword analysis.