Datathon-Network-Impact

Datathon Network Impact - Comparing the evolution of scholarly growth and Networking in attendees and presenters of SCCM datathon over the years from 2017 vs 2023.

Analysis Overview:

Comparing Diversity in SCCMDAT and SCCMPRES Datasets in the years 2017 vs 2023.

Datasets used:

1. For gender diversity analysis:

• "SCCMDATupdated_FINALANALYSISCSV.csv"

• "SCCMPRESupdated_FINALANALYSIS_csv.csv"

2. For Country diversity analysis:

• "fork_publication_summary (2)_countries_updated.csv" (for SCCMDAT)

• "PRESfork_publication_summary (2)_countries_updated.csv" (for SCCMPRES)

3. For Educational diversity analysis

• "finalcombanalysis_degree_comparison_updated.csv" (for SCCMDAT)

• "finalcombpresanalysis_degree_comparison_updated.csv" (for SCCMPRES)

Objective:

This analysis aims to compare diversity between two datasets, SCCMDAT and SCCMPRES, focusing on three key aspects:

1. Gender diversity

2. Country diversity (based on high-income vs. low-income countries)

3. Educational diversity

Methodology:

• Utilizes the Gini Impurity Index to quantify diversity

• Gini Impurity scale: 0 (complete homogeneity) to 1 (maximum diversity)

• Analysis conducted at the publication level for each dataset

• Comparison of results between 2017 and 2023 for both datasets (SCCMDAT and SCCMPRES)

Analysis Components:

1. Gender Diversity Analysis:

o Employed NLTK (Natural Language Toolkit) for gender prediction based on author names

o Developed a function to calculate Gini impurity for gender diversity

o Applied the function to SCCMDAT and SCCMPRES datasets for 2017 and 2023

o Created visualizations (box plots, histograms) to compare gender diversity across datasets and years

3. Country Diversity Analysis:

o Utilized datasets categorizing co-authors as from high-income or low-income countries

o Developed a function to calculate Gini impurity for country diversity

o Applied the function to both datasets for 2017 and 2023

o Generated visualizations (box plots, line plots) to illustrate country diversity trends over time

5. Educational Diversity Analysis:

o Incorporated datasets containing information on authors' highest educational degrees

o Developed a function to calculate Gini impurity for educational diversity

o Applied the function to both SCCMDAT and SCCMPRES datasets

o Produced visualizations (box plots, histograms, scatter plots) to compare educational diversity between datasets