Projects at Mayerfeld Consulting Data Analytics Practicum

These projects come from my Data Analytics Practicum with Mayerfeld Consulting, a remote program built around industry-style business case studies. The brief in every case was the same: start from a raw dataset and a business question, and end with something a stakeholder can act on.

E-commerce Performance Dashboards

I designed and built two Tableau dashboards on a multi-country e-commerce dataset covering roughly €3.4M in purchase value. The Revenue dashboard tracks category performance, acquisition channel mix across eight countries, and session trends by device; the Marketing Campaigns dashboard analyzes campaign spend against conversions and compares conversion and click-through rates by channel. Every chart is paired with a written takeaway – for example, that spending above the €28K median showed no conversion advantage, and that email marketing matched other channels’ 8-9% conversion rate at far lower cost.

Revenue dashboard: category, channel, and device performance across 8 markets
Marketing Campaigns dashboard: spend efficiency and channel conversion analysis
TAGS
Tableau (calculated fields, KPI banners, reference lines, layout) Data storytelling AOV / conversion rate / CTR Channel mix Campaign efficiency
WHAT I DID
Individually designed and built both dashboards, from data connection through chart design to the written insights.

Hypothesis Testing on National Survey Data

Working in a four-person team, I analyzed ten waves of the Young Life and Times survey (9,168 respondents) to test whether gender is associated with how important 16-year-olds in Northern Ireland consider their religious identity. After cleaning and recoding the data in Python – documenting every exclusion, from missing values to a question absent from an entire survey year – we ran a chi-square test of independence on the final sample of 7,804. The result was statistically significant (chi-square = 16.99, p = 0.002), and in the opposite direction to international patterns: boys, not girls, were over-represented at the strongest end of the scale.



Result

χ²(4) = 16.99
p = 0.00195
H₀ rejected at α = 0.05
n = 7,804

EXCLUSION LOG
Variable Code Reason N excluded
rsex 99 Missing 42
relidimp 9 Don't know - not analytically useful 19
relidimp 99 True missing 75
relidimp 999 Question not included in 2010 wave 786
relidimp 6 Qualitative opt-out, not part of ordinal scale 451
SAMPLE SUMMARY
N
Original sample 9,168
Total excluded 1,364
Final clean sample 7,804
of which: Male 3,171
of which: Female 4,633
Data cleaning log: every exclusion documented, from 9,168 raw respondents to the final analysis sample of 7,804. Swipe a table sideways on mobile to see all columns.
TAGS
Python (pandas) matplotlib Chi-square test of independence Hypothesis formulation Assumption checking UK Data Archive teaching dataset
WHAT I DID
Co-authored the analysis in a four-person team, working across data cleaning in pandas, the statistical test, and the written report.
hk-hande-logo