US Education Analysis
Overview
Motivation
Beginning my career in Education, I've learned to optimize resources for student success. While teaching, resources were plenty and results were reasonable. But outside of my bubble, how are other students performing? From a zoomed-out, birds-eye view, how is our nation performing with less, ample, and large sums of education funding?
Objective
Discover trends and patterns in US Education data and formulate meaningful insights.
Scope
The analysis covers Math and Reading Test scores (Grades 4 and 8) in all 50 States from 1992-2016.
The Process
Select & Clean Data
Research dataset with enough data and breadth for analysis. Import dataset into Python, and perform consistency checks to ensure data is ready for analysis (duplicates, null, and non-uniform data).
Develop & Test Hypothesis
Now that data is validated, brainstorm a hypothesis which could be tested via Python libraries (matplotlib, scipy, sklearn, numpy, json). Determine the hypothesis accuracy using Statistical knowledge.
Pivot
Because the hypothesis was incorrect, conduct a Cluster Analysis (k-means clustering) to find insight in the data's behavior. Use this insight to explain to stakeholders.
Stop & Visualize
Sifting through multiple tests and analyses, discern which visualizations would be most beneficial to tell the story.
Create Deliverables
Compile the relevant visualizations and tell the story.
Hypothesis
Greater Education Funding received per State will result in higher scores on Math and Reading Assessment Tests.
Tests & Analysis conducted
Get to Know Us
-
Linear Regression
-
Correlation Matrix Heatmap
-
Categorical Plot
-
Geospatial Map
-
Cluster Analysis (k-means clustering)
-
Time Series Analysis
Challenges
-
The initial hypothesis "failed," or was incorrect, so it was challenging to find any silver lining or alternative direction. Initially, it felt like the analysis was a failure and I needed to go back to the drawing board.
-
​​Resolution: Upon reflecting on the results, however, I found a purpose for this analysis behind this strange phenomenon - higher-funded States do not always result in better scores. Despite my hypothesis being incorrect, there was a reason to continue further.
-
-
The analysis results are not that simple.
-
​Resolution: More work needs to be done. There is plenty of "grey area" that needs to be discovered - more grade levels, more subjects, cause and effect from the pandemic (extending the time beyond 2020), budget allocation for each State, family's socioeconomic status, and location in urban, rural, or suburban neighborhoods.
-
Visualizations
The highest funded states are CA, TX, FL, NY, PA.
Summary of Cluster Data (TOP: Education funding, MIDDLE: 8th Math Test Scores, BOTTOM: 8th Reading Test Scores)
The highest funded states are CA, TX, FL, NY, PA.
Recommendations
-
Review the Low and Adequate Funded States’ Education budget plans – observe how these States intentionally spent for optimal performance.
-
Reassess High Funded States’ Education budgets – re-prioritize funds with academics as a priority.
-
Collect more up-to-date, diversified data (different grade levels and subjects) and analyze other variables affecting student performance.