Nikita Slavin,
cartographer
cartographer
How old is this house, Helsinki
A brief story about creating the Helsinki map of the ages of buildings
đ HOW COME?
The dogs guard selected maps of Russian cities from the project.
In every city Iâve lived in since March 2022, Iâve tried to create a building age map, by habit instilled by how-old-is-this.house (my rather successful project about the age of buildings in 75 Russian cities).
In Belgrade and my dearly beloved Tbilisi, there wasnât enough open data to make a good map.
I moved to Helsinki in August 2024. I gave it a try with age map - and it worked - open data is perfect, and I had enough resources and inspiration (I think it is a good sign đ)
So, I created a building age map with decent quality and great coverage.
In Belgrade and my dearly beloved Tbilisi, there wasnât enough open data to make a good map.
I moved to Helsinki in August 2024. I gave it a try with age map - and it worked - open data is perfect, and I had enough resources and inspiration (I think it is a good sign đ)
So, I created a building age map with decent quality and great coverage.
𧩠Data Sources and Processing
The process of searching, preparing, and processing data typically follows a same scenario: finding sources, evaluating their completeness and quality, prioritizing, and combining them.
If this process isnât automated but done manually, each time it becomes a new story, with its unique flavour and twistsâwhich is particularly enjoyable (for example, the story of the creation of the first map in the project â St. Petersburg, or Moscow map).
The core of the dataset was derived from Buildings in Helsinki from the City of Helsinki Urban Environment Division, which provided the majority of building polygons along with their key attributesâaddress, year of construction, and number of floors.
If this process isnât automated but done manually, each time it becomes a new story, with its unique flavour and twistsâwhich is particularly enjoyable (for example, the story of the creation of the first map in the project â St. Petersburg, or Moscow map).
The core of the dataset was derived from Buildings in Helsinki from the City of Helsinki Urban Environment Division, which provided the majority of building polygons along with their key attributesâaddress, year of construction, and number of floors.
The Buildings in Helsinki dataset overview with basic red-green colour ramp
OSM bridges
It might seem like the work is doneâand there wasnât much work to begin with.
But things arenât that simple: to make the dataset dynamic and engaging to explore, I try to enrich it with additional data as much as possibleâphotographs, architectural style information, architect details, and links to websites.
Before doing this, however, itâs necessary to add geometry for missing objects:
âin cities near water, bridges play an essential role, so I added them to the map by downloading all the bridges in Helsinki from OSM using Overpass API.
â just in case, I cross-referenced buildings in the Helsinki dataset with those extracted from OSM and added the missing onesâthere were only a few (likely due to philosophical or bureaucratic questions about what constitutes a âbuildingâ).
But things arenât that simple: to make the dataset dynamic and engaging to explore, I try to enrich it with additional data as much as possibleâphotographs, architectural style information, architect details, and links to websites.
Before doing this, however, itâs necessary to add geometry for missing objects:
âin cities near water, bridges play an essential role, so I added them to the map by downloading all the bridges in Helsinki from OSM using Overpass API.
â just in case, I cross-referenced buildings in the Helsinki dataset with those extracted from OSM and added the missing onesâthere were only a few (likely due to philosophical or bureaucratic questions about what constitutes a âbuildingâ).
Geocoded finna.fi API request in QGIS
Once the geometry was complete, it needed to be enriched with architectural attributes. For this, I used:
- The incredibly user-friendly and comprehensive API from Finna.fi: some records already had coordinates, but most of the ones of interest only had addresses, so the output had to be geocoded.
Wikidata points for Helsinki, red ones - with photo
2. Wikidata: a structured Wikipedia-family knowledge base, also a great data source.
3. Sometimes attributes were sourced from OSM.
4. An interesting source of information was found for four districtsâLauttasaari, Munkkiniemi, Pohjois-Haaga, and Laajasaloâbased on an inventory conducted by the Helsinki City Museum.
3. Sometimes attributes were sourced from OSM.
4. An interesting source of information was found for four districtsâLauttasaari, Munkkiniemi, Pohjois-Haaga, and Laajasaloâbased on an inventory conducted by the Helsinki City Museum.
List of dataset's fields
For each field in the final table, I defined a source priority order. For example: first the photo link is taken from Wikidata; if unavailable, then from the Finna API; if still unavailable, we rely on OSM.
Finally, the resulting dataset was compiled!
Finally, the resulting dataset was compiled!
At this point, I realized that Helsinki and the Helsinki metropolitan area are significantly different in geographical extent.
So, I repeated all the steps described in this chapter from the beginning.
So, I repeated all the steps described in this chapter from the beginning.
This extent of the dataset is so much more engaging and fun!
The next stage: evaluating the datasetâs adequacy, completeness, and overall quality.
The dataset contains 202,165 objects, of which 154,212 have a recorded construction year (~76%). However, for the central part of the city, this value reaches 99%. Based on my experience, these are solid numbers. More importantly, exploring the map creates a sense of cohesion and informativenessâa sign of successful data mining.
Absolute numbers for objects with photos, architects, or architectural styles arenât as impressive: 3,000â5,000. However, this covers a significant number of buildings in the historic centre, which is of the most interest.
The dataset contains 202,165 objects, of which 154,212 have a recorded construction year (~76%). However, for the central part of the city, this value reaches 99%. Based on my experience, these are solid numbers. More importantly, exploring the map creates a sense of cohesion and informativenessâa sign of successful data mining.
Absolute numbers for objects with photos, architects, or architectural styles arenât as impressive: 3,000â5,000. However, this covers a significant number of buildings in the historic centre, which is of the most interest.
The histogram which shows building counts by year reveals typical outliers:
â Rounded construction years (e.g., by decades) due to archival simplifications.
â A massive spike in 1900, likely representing 19th-century buildings recorded this way during archival digitization. While this is a common issue, it doesnât affect the spatial distribution of the dataset. Nevertheless, itâs worth considering when using the data.
Some other data checks highlights:
â checking on the oldest building: I've got a chapel in the Muuralan pappila complex. It seems a realistic result.
The oldest building in Helsinki city is the Sederholm house, which the maps proofs. But the dataset is created for the Helsinki Metropolitan area â and it appears the chapel near the Muuralan pappila church is the oldest building in the area, dated 1726 by the City of Helsinki Urban Environment Division.
â ensuring there are no buildings from the future (sometimes under-construction buildings are marked this way in OSM).
â adding missing information for key bridges and objects in the city centre by hand
As with any worthy cartographic project, the work is best done collaboratively. I asked Kirill Busygin-Tapani, a Helsinki tourist guide, to review the dataset. Many thanks for the help and advice!
Hooray! The dataset is ready!
It is freely available for any projects under the CC by-SA license on page "Dataset".
â Rounded construction years (e.g., by decades) due to archival simplifications.
â A massive spike in 1900, likely representing 19th-century buildings recorded this way during archival digitization. While this is a common issue, it doesnât affect the spatial distribution of the dataset. Nevertheless, itâs worth considering when using the data.
Some other data checks highlights:
â checking on the oldest building: I've got a chapel in the Muuralan pappila complex. It seems a realistic result.
The oldest building in Helsinki city is the Sederholm house, which the maps proofs. But the dataset is created for the Helsinki Metropolitan area â and it appears the chapel near the Muuralan pappila church is the oldest building in the area, dated 1726 by the City of Helsinki Urban Environment Division.
â ensuring there are no buildings from the future (sometimes under-construction buildings are marked this way in OSM).
â adding missing information for key bridges and objects in the city centre by hand
As with any worthy cartographic project, the work is best done collaboratively. I asked Kirill Busygin-Tapani, a Helsinki tourist guide, to review the dataset. Many thanks for the help and advice!
Hooray! The dataset is ready!
It is freely available for any projects under the CC by-SA license on page "Dataset".
đ DESIGN: COLOUR RAMP AND POSTER
Moving from data extraction and processing to the fun partâcartographic design.
I plan to use a trusted and beloved red-to-blue colour ramp from my previous city projects: it is highly distinguishable and effectively reveals the building structure. The red and yellow hues of earlier construction years intuitively evoke a sense of warm, incandescent lighting, while the blue tones correspond to LED lights.
I also want to tie the colour palette to the city by incorporating âits colours.â The designers of Helsinkiâs visual identity provided a palette that matches my impression of the cityâthank you!
I plan to use a trusted and beloved red-to-blue colour ramp from my previous city projects: it is highly distinguishable and effectively reveals the building structure. The red and yellow hues of earlier construction years intuitively evoke a sense of warm, incandescent lighting, while the blue tones correspond to LED lights.
I also want to tie the colour palette to the city by incorporating âits colours.â The designers of Helsinkiâs visual identity provided a palette that matches my impression of the cityâthank you!
Helsinki city colours, and colour ramp for the map inspired by them
While I couldnât exactly replicate these tones on the map, the key points align closely with the colours Brick, Summer, Coat of Arms, Engel, and Copper.
Iâve adjusted the distribution of the construction years across the colour ramp to make the visualization informative and visually appealing.
Iâve adjusted the distribution of the construction years across the colour ramp to make the visualization informative and visually appealing.
The next step is poster layout:
Usually a routine and enjoyable process, but it wasnât easy this time. The layout had to endure several test prints and plenty of feedback before finding the perfect size, map extent, and orientation.
Usually a routine and enjoyable process, but it wasnât easy this time. The layout had to endure several test prints and plenty of feedback before finding the perfect size, map extent, and orientation.
By the way, for this project, I used Affinity software instead of Adobe for the first time. While itâs not an ideal replacement, the cost-benefit tradeoff is fairâhighly recommend it to others.
The map on the poster is enhanced with elegant infographicsâa histogram displaying the distribution of buildings by construction year, which also doubles as the colour legend for the map.
The map on the poster is enhanced with elegant infographicsâa histogram displaying the distribution of buildings by construction year, which also doubles as the colour legend for the map.
The result is a quite nice posterâavailable for order in either a tube or frame through the map store.
đ CONTRIBUTION
Automated data processing is never perfect: errors occur, and attributes and geometries are sometimes incorrectly matched. Itâs normal for the dataset of Helsinki's building age map to have missing or inaccurate data.
The map is hosted on the Geosemantica platform, and the interface allows users to suggest corrections if they spot inaccuracies (use the "Edit" button in the top-right corner of a building's card).
Please contribute! After moderation, your edits will appear on the map, making it more accurate and comprehensive.
The map is hosted on the Geosemantica platform, and the interface allows users to suggest corrections if they spot inaccuracies (use the "Edit" button in the top-right corner of a building's card).
Please contribute! After moderation, your edits will appear on the map, making it more accurate and comprehensive.
đ THANKS
It was fun to do this project, I want to say thank you to:
- friends who put up with me and tested different parts and stages of the project
- Helsinki for being Helsinki