Question

It’s time to consider a topic to base your data story on. Consider the following questions to help guide you in your brainstorming process.

Which type of story (from the previous article) would you like to tell?

What topic do you find interesting or important that you’d like to bring awareness to by telling a data story? Feel free to list a couple of options at this point.

Conduct a bit of preliminary research. Is there a dataset available to help you tell your story? If not, could you create a survey and collect the data yourself?

Although this activity is done individually, you and your group will ultimately decide on the story you want to tell together.

Answer 1

Good — here’s a structured way to brainstorm and pick a data-story topic. I’ll propose types of stories, several topic options (with dataset availability and feasibility), and next steps you can take individually before discussing with your group.

1) Which type of story might you tell?
- Explanatory: show how something works or why a pattern exists (good for clear cause/effect or trends).
- Persuasive/advocacy: use data to convince readers to support an action or policy.
- Investigative: uncover surprising or hidden facts in messy data.
- Exploratory: present interesting patterns without a firm conclusion (good when you’ll be learning as you analyze).
- Predictive: forecast a future value or trend.
Pick the type that best fits your goal — e.g., persuasive for policy change, explanatory for clarifying trends.

2) Topic options (with data availability and feasibility)

A) Public transit ridership and recovery after COVID
- Story type: Explanatory or persuasive (e.g., argue for service adjustments or funding).
- Data: National Transit Database (NTD), local transit agency ridership reports, Google Mobility Reports, city open data.
- Feasibility: High — time-series data readily available; can map routes and compare pre/post-pandemic.
- Visuals: time series, maps, ridership by route, correlation with remote-work metrics.
- Survey needed? Not required, but a rider survey could add qualitative insights.

B) Student mental health and screen time
- Story type: Investigative/persuasive.
- Data: Some public sources (Pew, CDC Youth Risk Behavior Surveillance for high-level), but detailed local data likely requires a survey.
- Feasibility: Moderate — you can combine national datasets with a short class survey to get local context.
- Survey: 10–15 questions (screen time, sleep, mood scales, demographics). Aim for 50–100 responses.
- Ethics: anonymize responses and get consent.

C) Air quality and respiratory health (local focus)
- Story type: Explanatory/investigative.
- Data: EPA Air Quality System (AQI, PM2.5), state/local health department hospitalization/ER data or CDC WONDER, meteorological data from NOAA.
- Feasibility: Moderate — air quality data is public; health data may be aggregated or require requests.
- Visuals: maps, time series, correlation/regression analysis lagging AQ exposure vs. admissions.
- Survey needed? Not necessary, but interviews with local health professionals could add depth.

D) Food access / food deserts in our city
- Story type: Persuasive/explanatory.
- Data: USDA Food Access Research Atlas, Census ACS (income, car ownership), SNAP retailer locations, local store listings.
- Feasibility: High — datasets public; straightforward spatial analysis using GIS or basic mapping tools.
- Visuals: choropleths, distance-to-store heatmap, overlays with income/public transit.
- Survey needed? Optional (e.g., residents’ shopping habits, barriers).

E) Remote work trends and changes in commuting
- Story type: Explanatory/predictive.
- Data: American Community Survey (commute mode), Google Mobility, local employer/LinkedIn reports; you could also run a survey of classmates or local workers.
- Feasibility: High for regional/national trends; local prediction benefits from a targeted survey.
- Visuals: time series, cohort comparisons, model of future commute volumes.

3) Preliminary research and dataset access
- Start by identifying one or two topics you like. For each, do quick searches:
- Kaggle, data.gov, city open-data portals, Census/ACS, USDA, EPA, CDC WONDER, National Transit Database.
- Use keywords: “[city] open data + transit/ridership”, “EPA AQI data download”, “USDA food access atlas csv”.
- Download a small sample (one month or one year) to confirm fields and quality.
- If needed data isn’t public (e.g., local hospital admissions), check whether aggregated data is published or whether you can request it.

4) If you need to collect your own data (survey tips)
- Keep it short (10–15 questions), mix closed (Likert, multiple choice) and 1–2 open text questions.
- Ask about demographics minimally (age group, neighborhood) to allow analysis while protecting privacy.
- Provide consent statement and anonymize responses.
- Target sample size: 50–200 depending on classroom size and scope.
- Tools: Google Forms, Microsoft Forms, or Qualtrics (if available).

5) Tools and next steps
- Tools: Excel, Google Sheets for quick EDA; Python (pandas, matplotlib/Seaborn), R, or Tableau/PowerBI for visualization; QGIS or ArcGIS for spatial work.
- Next steps (individually, before group meeting):
1. Pick your top 1–2 topic choices and preferred story type.
2. Find one or two relevant open datasets and download a sample.
3. Sketch 2–3 questions you’d want the data to answer.
4. If you’ll need survey data, draft 8–12 potential survey questions.

Recommendation (quick): If you want high data availability and relatively quick analytics, go with either (A) public transit ridership or (D) food access. If you want a human-focused, persuasive story and don’t mind collecting data, pick (B) student mental health and run a short survey.

Tell me which 1–2 topics you prefer, and I’ll help you:
- locate specific datasets and links,
- draft a survey if needed,
- outline a simple story arc and visualization plan.