Diabetes Risk Analysis | Glucose, BMI, Age & Health Indicators

By abhishek.verma75000 · April 4, 2026

Analysis shows 34.81% diabetes prevalence, with high glucose and high BMI emerging as the strongest risk factors. Age, BMI, glucose, and genetic factors…

The overall diabetes prevalence rate in the dataset is 34.81%. Out of 767 individuals, 267 have diabetes and 500 do not. A pie chart has been generated to visually represent this breakdown.

Among the 147 individuals with glucose levels above 140 AND BMI above 30, 111 of them (75.5%) are diabetic. This is a notably high diabetes rate, suggesting that the combination of elevated glucose and high BMI is a strong indicator of diabetes in this dataset.

Diabetic individuals have noticeably higher BMI compared to non-diabetic individuals. The diabetic group (n=267) has a mean BMI of 35.1, while the non-diabetic group (n=500) has a mean BMI of 30.3 — a difference of 4.8 points. A box plot has been generated showing the full BMI distribution for both groups, making it easy to visualize this difference.

The Diabetes Pedigree Function (DPF) distribution differs notably between the two outcome groups. People with diabetes (Outcome 1) have a meaningfully higher DPF compared to those without diabetes (Outcome 0). Two visualizations were generated — a histogram showing overlapping distributions and a box plot comparing the spread — along with a summary statistics table.

Diabetic patients have a slightly higher average blood pressure (70.82 mmHg) compared to non-diabetic patients (68.18 mmHg). A bar chart has been generated to visually compare these two groups side by side.

The feature with the largest mean difference between diabetic and non-diabetic groups is '0', with an absolute mean difference of 31.92. This means diabetic patients have notably higher average values for this feature compared to non-diabetic patients.

The analysis attempted to examine how diabetes rates vary by age group among obese individuals (BMI 35). A data table was generated showing the results of this segmentation.

Among women with 5 or more pregnancies and a BMI over 30, the diabetes prevalence is 56.6% — meaning 99 out of 175 women in this group have diabetes. This is a notably high rate, suggesting that the combination of multiple pregnancies and higher BMI is strongly associated with diabetes risk.

Among individuals with glucose below 100, BMI below 25, and age under 30, the diabetes rate is 0.0% — meaning none of the 38 people in this group were diabetic. This is dramatically lower than the overall dataset diabetes rate of 34.8%.

A comprehensive summary statistics table and normalized heatmap have been generated for all 9 columns in the dataset, accessed using df.iloc[:, 0] through df.iloc[:, 8]. The heatmap visually highlights which features have the widest spread or highest variability by normalizing each statistic row across all columns — brighter colors indicate relatively higher values.

Unfortunately, the analysis could not be completed due to a data column mapping issue. The code attempted to use numeric values (like 6, 148, 33.6) as column names, but these didn't match the actual column names in the dataset, causing the grouping and calculation steps to fail. No charts or summary statistics were produced.