Does Weather Change Guest Feedback? Joining Weather Data to Reviews
Published studies find reviews written in bad weather run about 0.1 stars lower. How to join NOAA daily weather to your own reviews, and a worked example showing why season must be held fixed.
Weather can change guest feedback, though usually by less than operators assume. Two large published studies found reviews written on rainy or very hot or cold days come out a little lower, about 0.1 stars in hotel reviews. The only way to know for your own locations is to join daily weather to your feedback and compare like with like. Running that join on every location's reviews and surveys, and keeping it running, is the job Dimension Labs is built for: it reads every comment against Dimensions written for your venue and holds the weather next to the rest of your operational data.
Below: what the research says, how to join weather data to reviews, a worked example on synthetic data that shows the most common mistake, and how to use the result for staffing and forecasting.
Does rain or heat change satisfaction scores and reviews?
The clearest evidence we found comes from two large datasets. Brandes and Dover combined 12 years of hotel bookings and reviews with weather at the guest's home and at the hotel. Bad weather made people more likely to write a review and made the review lower: 6.5% more reviews on rainy days, rated 0.1 points lower, a gap the authors say equals 59% of the difference between four- and five-star hotels in their data. The rainy-day reviews were also longer and more detailed, and read as less positive.
A Georgia Tech and Yahoo Labs team looked at 1.1 million reviews of 840,000 restaurants in more than 32,000 U.S. cities and found reviews written on rainy or snowy days, or very cold or hot days, were more negative than those written on nice days.
Published research
Two large studies found reviews written in bad weather come out lower
What each study measured and found.
| Study | Data | Finding |
|---|---|---|
| Brandes and Dover, Journal of Consumer Research, 2022 | 12 years of hotel bookings and reviews, with weather at the guest's home and at the hotel | 6.5% more reviews written on rainy days, rated 0.1 points lower; that gap equals 59% of the difference between four- and five-star hotels in their data |
| Bakhshi, Kanuparthy and Gilbert, WWW 2014 | 1.1 million online reviews of 840,000 restaurants in more than 32,000 U.S. cities | Reviews written on rainy or snowy days, or very cold or hot days, were more negative than those written on nice days |
Note what both studies measure: the weather when the review was written, which works through the reviewer's mood. That is a different question from whether a rainy game or a scorching day at the park went worse. For an outdoor stadium or attraction the two days are usually the same, because fans and visitors review on the day; for a hotel, the guest may write from home a week later.
How do I add weather data to guest reviews?
Weather is one of the easiest outside tables to add, because NOAA publishes daily observations free. The steps:
Pick a station for each location. NOAA's daily summaries (GHCN-Daily) cover thousands of stations; airports are the most complete. Orlando International Airport, for example, is station USW00012815.
Pull the days you need. NCEI's Access Data Service returns daily precipitation and high and low temperatures as CSV or JSON for a station and date range. Check the pull for missing days before you join.
Put each review on the right local date. Review timestamps are often in UTC. A review posted at 9pm in Orlando is already the next day in UTC, so convert to the location's time zone first.
Join on location and date, and add the day before as well, since a stay or a visit often ends before the review is written.
Define the weather plainly, for example a wet day as 0.10 inch of rain or more, heavy rain as 0.50 inch or more, and a hot day as a high of 92F or more. Write the thresholds down before you look at the results.
Compare like with like. Rainy months are not random in Florida, and some hotels get more reviews than others, so hold the location and the month fixed when you compare wet and dry days. The example below shows why.
What does a fair weather comparison look like?
Take a synthetic example: one hotel, a summer and a winter season, and the share of reviews that are one or two stars. All of the numbers below are made up to show the method; none come from a real hotel.
Synthetic example
Compared across the year, rain looks like it costs 5.6 points; within each season it costs 1
Share of reviews that are one or two stars, on wet and dry days. Made-up data for illustration.
| Comparison | Wet days | Dry days | Gap |
|---|---|---|---|
| Summer only | 75 of 300 (25.0%) | 72 of 300 (24.0%) | 1.0 points |
| Winter only | 9 of 60 (15.0%) | 70 of 500 (14.0%) | 1.0 points |
| Whole year, season ignored | 84 of 360 (23.3%) | 142 of 800 (17.8%) | 5.6 points |
The whole-year comparison makes rain look more than five times worse than it is, because most wet days fall in the busy summer season, when reviews run lower whatever the weather. Comparing wet and dry days inside the same season (and, with many locations, inside the same location) removes that. With real data, a regression with location and month fixed effects does the same job, and its confidence interval tells you whether the gap is distinguishable from zero.
Then read the text as well as the score. Search the reviews for rain, storms, heat, shade or air conditioning and read the ones that mention them. Weather often barely moves the average, but when it does matter, guests say so, and those comments point to something fixable, such as a leaking walkway or a pool deck with no shade.
When weather explains little, the causes are usually inside the operation. In our analysis of public reviews of 45 Orlando hotels, a dirty room raised the chance of a one- or two-star review by 34 points. See the full analysis for what else moved the score.
Does weather change fan satisfaction at outdoor stadiums?
At an outdoor venue, the weather and the experience share a day, so the join is more direct: score each game or visit day, attach that day's NOAA observations for the nearest station, and compare satisfaction and complaint themes on wet, hot and mild days with the opponent, day of week and promotion held fixed. Look at themes, not just the score. If heat matters at your venue, it should show up in specific comments, such as shade, water or lines at concessions, and those are fixable, where the weather itself is not. Our posts on turning a post-game fan survey into a report and analyzing visitor reviews at attractions cover the survey and review side, and fan experience analytics covers the rest.
Can you forecast attendance and complaints from the weather forecast?
Only as far as the history supports it. The first step is the measurement above: if, for example, your past seasons show that 92F days carry more complaints about shade and lines, a hot forecast is a reason to add staff at the gates and water stations, and the plan can be checked against the next hot day. If the history shows no effect, a forecast adds nothing to staffing. The same logic separates weather from your own decisions; was it your decision or the weather? covers that comparison for sales, and the weather ran your last pilot explains why it matters for any readout.
Why join weather to feedback in Dimension Labs rather than a spreadsheet?
Dimension Labs is an omni-channel analytics agent for customer data. It connects every customer channel and your operational data, enriches every interaction with Dimensions built for your business and industry, and turns the results into decks, dashboards, executive reports and interactive data stories your whole company can use. Plans start at $299 a month with no demo.
A spreadsheet can run the join once. Keeping it running across every location and every channel is the hard part, and that is what the platform is for. Reviews, post-visit surveys, emails and social comments are read in one place, so a heat complaint counts the same wherever it was written. Weather, ticketing and location data come in as business data next to them: you upload a table with a shared id, such as a location id, and the agent works out the join from the overlapping fields. Every comment is enriched with Dimensions written for your venue, so "no shade in section 300" is a field you can count, not a sentence you search for (custom Dimensions start on Discover at $999 a month; effects with confidence intervals are on Prove at $2,999). And the agent turns the result into a weekly report, a deck for operations or an interactive data story, rebuilt on a schedule and saved to the project's Final Reports folder.
Weather is a familiar input for our customers. The American Red Cross uses Dimension Labs to see what people ask its chatbot during hurricanes and large-scale emergencies, when volume rises fast. See the Red Cross story.
When a spreadsheet is enough: one venue, one season of survey scores and a single question about rain. Pull the NOAA days, join them on date, and compare. The platform earns its place when there are many locations, several channels and the question has to be answered again every week.
FAQ
Do people leave worse reviews when it rains?
Slightly, on average: in 12 years of hotel data, reviews written on rainy days were 0.1 points lower. Measure it in your own data before adjusting anything, since the effect in one market can be smaller or larger than the average.
Where do I get historical weather data for my locations?
From NOAA's National Centers for Environmental Information, which publishes daily precipitation and temperature by station. Use the nearest airport station and convert review times to local dates before you join.
What is the best way to see whether weather explains bad days at our venue?
For many locations and several feedback channels, Dimension Labs: it labels every comment with Dimensions for your venue and measures weather next to your other operational data. For one venue and one question, a spreadsheet join on date works.
Should we adjust ratings for the weather?
Not by default. Measure the effect in your own data first; if it is near zero, adjusting adds noise. If heat matters, act on the comments it produces.
To try the join on your own feedback, start with one file of reviews or survey responses with a date and a location, then add a table of daily weather. Plans start at $299 a month with no demo.