Diabetes Risk Factors Analysis 2026: Key Findings from Healthcare Data

By abhishek.verma75000 · June 13, 2026

This analysis examines eight health metrics to understand which factors most strongly predict diabetes. Glucose emerges as the dominant indicator—diabetic…

Yes, higher number of pregnancies is associated with increased diabetes prevalence. The data shows a clear upward trend — women with 0 pregnancies have a 34.2% diabetes rate, while those with 9+ pregnancies have a 55.8% diabetes rate. A bar chart has been generated to visualize this trend across pregnancy groups.

All 8 health metrics were analyzed for their correlation with diabetes Outcome. Glucose is by far the strongest predictor with a correlation of 0.47, followed by BMI (0.29) and Age (0.24). A horizontal bar chart and a full correlation heatmap were generated to visualize these relationships clearly.

Diabetes prevalence rises sharply with increasing BMI. The highest rates are found in the obesity categories (BMI 30+), with Obese III (BMI 40+) topping the list at 56.1%. A bar chart and data tables have been generated to visualize these findings clearly.

Diabetic patients have significantly higher average glucose levels (141.26) compared to non-diabetic patients (109.98), a difference of 31.28 units. Two visualizations were generated: a bar chart comparing the averages side-by-side, and a box plot showing the full glucose distribution for each group.

The analysis reveals how diabetes prevalence changes across different age groups, shown through two interactive charts and supporting data tables. A bar chart displays the percentage of diabetic patients in each age group, while a grouped bar chart compares the count of diabetic vs. non-diabetic patients side by side.

Diabetic patients have significantly higher average glucose levels compared to non-diabetic patients. Non-diabetic individuals average 110.64 mg/dL, while diabetic individuals average 142.32 mg/dL — a difference of 31.68 mg/dL, or about 28.6% higher. Two visualizations were generated: a bar chart comparing average glucose by outcome and a box plot showing the full distribution of glucose levels for each group.

Diabetic patients have a noticeably higher DiabetesPedigreeFunction compared to non-diabetic patients. The mean score for diabetic patients is 0.5505, versus 0.4297 for non-diabetic patients — a difference of about 28.1%. This suggests that individuals with diabetes tend to have a stronger genetic or family history predisposition to the condition. A box plot and histogram were generated to visually compare the distributions between the two groups.

The analysis successfully segmented patients into four BMI categories and produced a grouped bar chart comparing diabetic vs. non-diabetic counts, along with data tables summarizing each group's diabetes prevalence, average glucose, and average age.

The IQR method was applied to BloodPressure, Insulin, and SkinThickness columns. Two bar charts and two boxplots were generated to visualize the findings, along with detailed summary tables. BloodPressure had the most outliers at 45 records (5.86%), with valid IQR bounds of 35–107. Insulin showed 34 outliers (4.43%) but is heavily impacted by zeros — nearly half the records (48.7%) have a value of 0, which skews the IQR calculation significantly. SkinThickness had only 1 outlier (0.13%), but 29.56% of its values are zero. These zeros almost certainly represent missing data rather than true medical measurements, and should be treated accordingly before further analysis.

A Pearson correlation heatmap has been generated for all 9 numeric columns in the diabetes dataset (Pregnancies, Glucose, BloodPressure, SkinThickness, Insulin, BMI, DiabetesPedigreeFunction, Age, and Outcome). The heatmap uses a Red-Blue color scale where red indicates positive correlations and blue indicates negative correlations. Correlation tables are also available for detailed exploration.

A comprehensive side-by-side summary statistics table was generated comparing all 8 numeric features between diabetic (Outcome=1) and non-diabetic (Outcome=0) patients. Two bar chart visualizations were also created — one showing mean values for each group, and another showing the percentage difference in means. Diabetic patients consistently show higher averages across nearly all features, with Pregnancies (47.5% higher) and Insulin (45.9% higher) showing the greatest divergence, followed by Glucose (28.4%) and DiabetesPedigreeFunction (28.1%). Blood Pressure shows the smallest difference at just 3.9%.