Project Vogue: Engagement Analysis

Project Vogue

Power BI — an interactive report analysing participation, retention and satisfaction across 16 Yoruba cultural heritage sessions delivered by MYEC between October 2024 and April 2025.

Tools & Skills

  • Power BI data modelling & DAX measures
  • Cross-filtering & drill-through navigation
  • Data cleaning, deduplication & quality auditing
  • Categorising free-text survey responses

Approach

517 exit-survey responses across three programme events were cleaned in Excel, then modelled in Power BI. Participant names were reconciled to a single canonical form, gender and programme values standardised, dates verified against the delivery calendar, and free-text responses grouped into consistent themes so they could be counted rather than merely read.

The report runs to four pages: a Summary carrying programme-wide KPIs, plus a drill-through page for each event with its own participation, gender, feedback and referral views.

Data Quality Findings

Two issues surfaced during preparation, both of which would have produced plausible-looking but wrong results. Participants wrote their names differently at each session, so 169 distinct name strings resolved to just 66 real people; taken at face value the register would have reported 2.6 times more unique participants than the programme actually reached. Separately, eight of fifteen sessions carried a 2024 year stamp for dates that fell in 2025, which would have placed Rehearsal six months before Storytelling began.

Key Findings

  • Retention, not reach. 560 attendances came from just 66 people, roughly 8.5 visits each. Half the cohort attended ten or more sessions across seven months.
  • The clearest gap. 17 participants attended exactly once and never returned, which is also why the 9.45 referral rating should carry a caveat.
  • Satisfaction is unusually clean. Across 517 responses on a 0 to 10 scale, the lowest score recorded was 8. There are no detractors in the dataset at all.
  • Material culture leads. Traditional outfit styles drew 116 mentions, ahead of proverbs at 78. Tangible heritage lands more memorably than oral heritage.
  • Food and friendship drive return. Catering and making new friends drew 109 mentions each, both ranking above requests for more cultural content.

Report Interactivity

Two mechanisms compound. Cross-filtering means clicking any element filters every other visual on the page instantly. Drill-through moves to a dedicated page per event, which can then be cross-filtered again. Filtering Storytelling to its final two sessions drops "Next workshop" from 56 mentions to zero, since once the last session had passed nobody asked for another one.

Outcome

Four recommendations tied to specific figures, including following up the single-visit group, scaling Oral History deliberately from its ten-attendee pilot, protecting the catering budget as programme design rather than a discretionary extra, and reporting reach and cultural contact as two distinct measures so funders are not given the wrong number.