matplotlib-pharmaceutical-analysis

Analysis using matplotlib and pandas of pharmaceutical data

Getting Started

Repository contains:

• A jupyter notebook file with all analyses and summary insights (pharmaceutical-analysis.ipynb)
• Two csv files used in the analysis (Mouse_metadata.csv and Study_results.csv)

The jupyter notebook contains additional details of the analysis

Features

Cleaning -

• Code will merge two csv files into a single dataframe (mouse_study_df)
• Finds if any mice in the study have identical timestamps
  • Creates two new dataframes:
    • Contains only the duplicate data for reference (duplicate_mouse_df)
    • Excludes the duplicate data, used for further analyses (mouse_study_clean)

Summary Statistics -

• Tables of summary statistics:
  • Mean, median, variance, standard deviation, and sem (standard error of the mean)
• Includes two methods:
  • Pandas approach
  • Aggregation method using numpy and scipy

Bar and Pie Charts -

• There are two versions of each chart:
  • One uses pandas.plot and the other uses matplotlip.pyplot
• Bar charts show total count of measurements for each of the 8 treatment groups
• Pie charts show the distribution of female and male mice in the study

Quartiles, Outliers, and Boxplots

• Qunatitative investigation of potential outliers based on final measurement of tumor volume (mm3)
  • Analysis looks at tumor volume at last timepoint for mouse per 4 treatment groups of interest
  • Includes count of potentual outliers
  • Box plot of final tumor volume with outlier noted as a blue star

Line and Scatter Plots

• Line chart of a single mouse from the study
  • Mouse ID is sampled randomly from a subset based on survival in study and treated with Capomuli
  • Chart is number of days by tumor volume
• Scatter plot is average tumor volume by mouse weight for Capomuli regimen

Correlation and Regression

• Correlation coefficient calculated using pearson's
• Linear regression and r-squared calculated
  • The linear regression line equation is included in a variation of the scatter plot from the prior step
    • The x-axis and y-axis were changed to represent the independent and dependent variables in the study

Licensing by:

The code in this project is licensed under MIT license.