Market analysis · Airbnb
A Power BI dashboard to read the market beyond price.
I analysed 6,030 Airbnb listings in the Basque Country, starting with territory rather than price.
01 · The starting point
Understanding a market before looking at price.
I wanted to analyse what the Airbnb market in the Basque Country is really like, using public data. Instead of stopping at the average price, I explored how supply changes with territory, type of accommodation and host profile.
This project brings the whole process together: data cleaning, modelling, analysis and a Power BI dashboard designed to answer real questions.
02 · The goal
Reading the market by territory before price.
The first thing I did was understand the size of the market. Before calculating prices or trends, I needed to know how many properties there were, how many hosts took part and how supply was spread.
That initial view helped me interpret the rest of the analysis better.
listings analysed
locations in 3 provinces
different hosts
03 · My approach
The tourism market is spatial before it’s statistical.
When I put the data on the map, one of the first important differences appeared: supply wasn’t evenly distributed.
Some areas had a high concentration of properties, and others much less activity. Seeing that distribution before getting into prices changed the way I read the market.
That’s why I decided to build the analysis starting from territory, not from the average.
This is how I summarised the market’s distribution.
Budget
Under €50/night
6.67%
Mid-range
€50 – €149/night
54.73%
High
€150 – €299/night
28.46%
Luxury
€300/night or more
10.15%
04 · Data model
I built a star schema to analyse the market in context.
After cleaning the data, I designed a star schema in Power BI, organising the information into a fact table and several connected dimensions.
This structure makes the analysis more consistent, makes calculating metrics in DAX easier and lets you move between territory, accommodation and time without losing the context of each figure.
Visualisation in Power BI
From the summary to the map without losing the filter.
I designed the dashboard so anyone could start with an overview and then dig into a specific area.
Filters keep their context across pages, so you can move from an executive summary to a detailed map without losing the thread of the analysis.
The signal that didn’t show up in the market average.
31 hosts (0.8%) — 15.2% of all supply.
The market average was hiding a very clear concentration.
Analysing how listings were distributed, I found that just 31 hosts, around 0.8% of the total, managed more than 15% of all supply.
That figure changed how I read the market: not all hosts take part in the same way. Some manage far more than the rest, something that only appeared when crossing territory, type of accommodation and the distribution of listings.
Key findings
Four things the average price doesn’t explain.
Geographical concentration
Mapping the properties, I saw that activity wasn’t spread evenly. Some areas concentrated a large share of supply, while others had a much smaller presence.
Professionalisation of supply
Analysing the hosts revealed an important difference between those who manage a single property and those who run several listings.
Dominant type
Most of the supply is entire homes, although the proportion changes with the territory.
Barely visible niches
Crossing variables revealed small segments that go unnoticed when you only look at the market average.
Tools
Tools I used.
Power BI
Modelling, visualisation and dashboard interaction.
DAX
Calculating metrics and measures.
Star schema
Organising the data model.
Power Query
Cleaning and transformation.
Dimensional model
Relationships between facts and dimensions.
Editorial visualisation
Design built to make reading easier.
What I take away from the project
The model found two real errors.
During the process, two problems came up that could have changed the reading of the analysis if they hadn’t been fixed.
The first had to do with duplicate records that inflated some metrics.
The second appeared when reviewing the relationship between hosts and listings, where part of the market’s concentration stayed hidden if only general aggregations were used.
Solving both forced me to validate the model before building the dashboard.
It was a good lesson: a visual can look good and still rest on a wrong calculation.
From a loose price to a market with structure.
This project sums up how I approached a complete analysis: from preparing the data to building a dashboard designed to answer business questions.
Interactive dashboard published in the Power BI Service
Full Power BI portfolio