Data Analysis in Healthcare

A lot of data is generated and used in the healthcare industry. This data can come from electronic health records, lab notes, imaging, financial reports, research studies and more. Analysis of this data is a really powerful tool that helps with decision making, finding early signs of illness, improving therapies and improving the patient's experience.

Different Types of Data Analysis

There are 4 different types of data analysis. The first type is descriptive analysis. This type of analysis is related to summarizing, visualizing and presenting the data so that it can be understood and processed by others. This can involve looking at ER wait times, how many people visited a hospital in one day and statistics about disease outbreaks. Descriptive data analysis is a more surface level analysis that allows people to see patterns and trends.

The second type of data analysis is Diagnostic Analysis, which focuses on explaining why some trends and patterns exist. This could involve explaining why a patient was hospitalized or why the patient is experiencing certain symptoms. To do this type of analysis people usually identify anomalies and questions based on findings from descriptive analysis, and explore data to find the cause behind those anomalies.

The third type of data analysis is called predictive analysis. This analysis uses past data to predict future outcomes of decisions. It could predict the progression of a disease in a patient so that doctors can update the patient's treatment plan. It could also predict a good patient population for study groups in clinical trials to improve the reliability and efficacy of the trial.

Prescriptive analysis focuses on looking at past data and giving healthcare professionals instructions and advice with certain decisions. For example, prescriptive analysis can offer treatment plans for patients.

How Data is Explored and Visualized

There are many different ways that data can be analyzed, visualized, and explored. Data can be summarized using statistics such as mean, median and standard deviation. The shape and distribution of data can be seen using box plots, histograms and density plots. Whereas scatterplots and correlation matrices can be used to find correlations between data.

Some powerful tools for data visualization are graphs are charts, like line graphs, scatterplots, bar graphs and heat maps. Visualization helps understand the data and identify anomalies, trends and patterns. Geospatial analysis is also another interesting example of data visualization. It uses maps and locations to represent data, giving useful information about the health of different populations, disease outbreaks, access to health resources and more. Imaging like MRIs, CT scans and X rays are also examples of data visualization that healthcare professionals look at to make diagnoses and treatment plans. Visualization is also used for visualizing machine learning model predictions. Some visualization tools for this are decision trees, feature significance plots, and receiver operating characteristic curve.

Overall, data analysis is very important in healthcare as it assists physicians with decision making, helps with diagnoses and overall improves the patient's care.

Project Description

For this week, I used the same dataset as last week and visualized the data to try and find patterns and trends. I used matplotlib to visualize the data.

Kaggle link: https://www.kaggle.com/datasets/thedevastator/cancer-patients-and-air-pollution-a-new-link

Data Analysis Code