# Northside Community Café: customer survey findings

Example project: the café, people and data are invented to show what a Helpal delivery contains.

## Your questions and short answers

1. **Are weekday or weekend visitors more satisfied?** Weekday visitors. 66.2% rated their visit 4 or 5 out of 5, against 33.3% at weekends: a gap of 32.8 percentage points (95% confidence interval 22.1 to 43.5).
2. **Does waiting time change satisfaction?** Yes, strongly. 80.7% of people who waited 5 minutes or less were satisfied, falling to 18.8% for waits of 16 minutes or more (Spearman correlation −0.48, n = 350). Weekend visitors waited longer (median 12 minutes against 7), which explains much of the weekend gap: among people who waited 6–10 minutes, 61% of weekday and 50% of weekend visitors were satisfied.
3. **What do comments mention most?** Coffee and food (37.4% of comments) and friendly staff (32.6%). Seating and space came up in 19.8%, price in 14.4%, and waiting or queues in 9.3%.

## The data

412 responses were supplied. After removing 12 duplicate submissions, 400 remain: 368 with a usable rating, 350 with both a rating and a waiting time, and 334 with a comment. Every change is listed in `changes.md`.

## Table 1. Satisfaction by day of visit

Ratings run from 1 to 5; "satisfied" means a rating of 4 or 5.

| Day | Rated | Mean | Median | Satisfied | Median wait |
|---|---|---|---|---|---|
| Weekday | 263 | 3.79 | 4 | 66.2% | 7 minutes |
| Weekend | 105 | 3.19 | 3 | 33.3% | 12 minutes |

## Table 2. Satisfaction by waiting time

| Wait | Rated | Mean | Satisfied |
|---|---|---|---|
| 0–5 minutes | 88 | 4.11 | 80.7% |
| 6–10 minutes | 148 | 3.70 | 59.5% |
| 11–15 minutes | 98 | 3.16 | 34.7% |
| 16 minutes or more | 16 | 2.69 | 18.8% |

Satisfied, by day, within the same waiting time: 0–5 minutes, weekday 80% of 80 and weekend 88% of 8; 6–10 minutes, 61% of 122 and 50% of 26; 11–15 minutes, 48% of 46 and 23% of 52; 16 minutes or more, 100% of 2 and 7% of 14.

## Table 3. What comments mention

334 responses included a comment. Comments were coded by the keywords agreed in the scope; one comment can mention several themes.

| Theme | Comments | Share |
|---|---|---|
| Coffee and food | 125 | 37.4% |
| Friendly staff | 109 | 32.6% |
| Seating and space | 66 | 19.8% |
| Price | 48 | 14.4% |
| Waiting and queues | 31 | 9.3% |

## Methods

Descriptive summaries for each group. The weekday–weekend difference in the share satisfied has a 95% confidence interval from the normal approximation for two proportions. The link between waiting time and rating uses Spearman's rank correlation, because ratings are ordered categories rather than measurements. Comment themes use keyword coding.

## Assumptions and limitations

- This is a survey of people who chose to answer, not of every visitor.
- Waiting times are what people reported, often rounded.
- The results show an association. Longer waits go with lower ratings, but busy periods may also differ in other ways, so this does not prove that waiting causes lower satisfaction.
- Some groups are small: 16 ratings at 16 minutes or more, 8 weekend ratings at 0–5 minutes and 2 weekday ratings at 16 minutes or more. Treat those percentages as rough.
- Keyword coding is approximate. It misses comments that use other words, and it does not capture tone: "coffee fine but a bit pricey" counts for both coffee and price.

## How to reproduce

Run `python analysis.py cafe-survey-original.csv`. It prints Tables 1–3 and writes `cafe-survey-checked.csv`. It needs only Python 3.8 or newer.