Descriptive Statistics Explorer for Survey Data
Build an R script that turns raw survey answers into a clear summary: for every question, it shows the right kind of summary (average, median and spread for number questions such as age or a 1 to 10 rating, and counts and percentages for multiple-choice questions), draws a matching chart, and reports how many people skipped each question. Everything is combined into a tidy HTML report you could share with the people who ran the survey. It's a good first data project in R: you'll learn importing data with readr, cleaning and summarizing with dplyr, choosing the right summary and chart for each kind of data, ggplot2, and R Markdown.
- Estimate
- ~7.5h
- Steps
- 5
- Completed by
- 0
- Proposed by
- codeseed.app
dplyr · ggplot2 · readr
Project roadmap
- 01
Import and inspect the data
~1.5hFind a survey dataset (for example a public survey on Kaggle, or create a small one yourself) and load it with readr::read_csv, setting column types where they're guessed wrong. Use glimpse() and summary() to understand it, and count missing answers per question. Turn text answers with a fixed set of options into factors with a sensible order (for example Never, Sometimes, Often). Done when: every column has the right type, and missing answers are counted.
- 02
Summarize numeric questions
~1.5hFor each numeric question, compute the number of answers, mean, median, standard deviation, minimum and maximum with dplyr (summarise with across), ignoring missing values. Present them in one tidy table with a row per question, rounded sensibly. Done when: the table covers every numeric question with correct values.
- 03
Summarize categorical questions
~1.5hFor each multiple-choice question, count the answers and compute percentages with count() and mutate(), keeping the options in their logical order and showing skipped answers separately. Optionally compare answers between two groups of respondents (for example by age group). Done when: each categorical question has a frequency table that adds up to 100%.
- 04
Draw the charts
~1.5hWith ggplot2, draw a histogram for each numeric question and a bar chart of percentages for each categorical question, with clear titles (the question text), axis labels and a consistent style. Write a function that picks the chart type from the column type, so all charts come from a loop. Done when: every question has a suitable chart saved as a PNG.
- 05
Build the report
~1.5hCreate an R Markdown (or Quarto) report with an overview (number of respondents and completion rate), then a section per question with its summary table, chart and one or two sentences describing what stands out. Knit it to HTML. Done when: the HTML report shows every question clearly and can be understood without reading any code.
Ready to build this?
Get a GitHub repo and start building. Your AI reviewer checks each step as you go.
Tech stack