Cluster Analysis of districts by immunisation coverage and poverty rate Based on DBSCAN

Project info

• Field of studyPublic Health and Epidemiology
• LevelMaster's thesis (S2)
• Method typeClustering
• Estimated build time± 5 working days
• Toolsscikit-learn DBSCAN, NearestNeighbors, make_moons
• PriceUS$22 - US$144

Research goal

Apply DBSCAN to synthetic data to cluster districts by immunisation coverage and poverty rate and interpret each group's characteristics.

Variables

Example result

• 2Clusters found
• 6.667Noise points %
• 0.624Silhouette (non-noise)
• 0.311eps
• 5min_samples
• 0.43ARI vs. true groups

DBSCAN (eps 0.31, min_samples 5) finds 2 arbitrarily shaped clusters on District and Immunisation coverage, with 6.7% of points flagged as noise.

Computed from illustrative synthetic data, not from a real study.

About DBSCAN

DBSCAN groups points by density, so it finds arbitrarily shaped clusters and flags noise. It suits spatial data or data with outliers and non-spherical clusters.

Steps

1. 1Standardise features
2. 2Set min_samples from the data dimension
3. 3Choose eps with the k-nearest-neighbour distance plot
4. 4Run DBSCAN
5. 5Count clusters and noise points
6. 6Plot the result with noise marked

Assumptions to check

• Cluster densities are roughly uniform
• eps and min_samples are chosen carefully
• Features are comparably scaled since distance is used

Reading and reporting the results

Report eps, min_samples, number of clusters, and noise percentage, then discuss parameter sensitivity. Compare with K-Means to show the advantage on irregular cluster shapes.

Example Python code

A short starting script with this project's variable names. The delivered version is adapted to your data and fully commented.

import numpy as np
import matplotlib.pyplot as plt
from sklearn.datasets import make_moons
from sklearn.cluster import DBSCAN
from sklearn.neighbors import NearestNeighbors
from sklearn.preprocessing import StandardScaler

X, _ = make_moons(n_samples=400, noise=0.07, random_state=17)
X = StandardScaler().fit_transform(X)
immunisation_coverage = X[:, 0]
poverty_rate = X[:, 1]

k = 5
dist, _ = NearestNeighbors(n_neighbors=k).fit(X).kneighbors(X)
kdist = np.sort(dist[:, -1])

db = DBSCAN(eps=0.25, min_samples=k).fit(X)
district = db.labels_
n_clusters = len(set(district)) - (1 if -1 in district else 0)
print("Clusters:", n_clusters)
print("Noise points:", int((district == -1).sum()), "of", len(district))

fig, ax = plt.subplots(1, 2, figsize=(10, 4))
ax[0].plot(kdist); ax[0].set_title("k-distance plot"); ax[0].set_ylabel("distance")
ax[1].scatter(immunisation_coverage, poverty_rate, c=district, cmap="viridis", s=15)
ax[1].set_title("DBSCAN (noise = -1)")
plt.tight_layout()
plt.show()

Packages and prices

Choose how much help you need. Each higher package includes everything in the lower ones. Prices exclude 11% VAT.

• Payment: 50% down payment before work starts, balance after the deliverables are handed over
• Tax: Prices exclude 11% VAT
• Data and academic integrity: Results are computed from your own research data; you write and stand behind your manuscript, we help with the analysis and your understanding of it
• Revisions: Up to 2 revisions are included in the price (Package 3: 3 rounds)

Frequently asked questions