The Curio Tree Dataset

This dataset contains much of the tree inventory, images and stories data that was collected on the Curio platform before it was sunset. The data was extraced from a number of database tables and includes;

• The inventory details of 2.5 millions trees from locations across the globe (location, species, diameter at breast height (DBH), height, vitality etc, where available)
• 27,288 images of trees that were uploaded onto the platform by our community and linked to individual trees and their species information etc.
• Notes (stories), tags and conversations linked to trees.

Dataset Description

Curio was an environmental education and outreach platform that was predominantly focused on urban forestry. It connected the various stakeholders involved in the management of urban forestry with the public and importantly made all data uploaded via its web and mobile apps publicly available. The platform was live from March 2016 until August 2023 when the maintainence overheads made its ongoing availability infeasible. Curio was supported in its early stages by two European Space Agency projects, through the New Commons and Curio Canopy. A sense of the platform and how it worked can be found via the videos on its supporting youtube channel

This repository contains much of the tree inventory, images and stories data that was collected on the platform via our community, projects we helped support and open data tree inventories we uploaded onto the platform. We are keen to make this data available for research purposes in the hope it might be of benefit to others and to further the efforts of our community.

We have endeavored to name as many of those great projects and data sources that were hosted on the Curio platform in the attribution section below. If there are any omissions or errors please contact us.

A related project involved generating a high resolution map of tree canopy cover for the Greater London Authority. Details of that project and dataset can be found on the London Datastore Curio Canopy page.

• Curated by: Breadboard Labs
• License: cc-by-nc-4.0

Dataset Sources and Attribution

Many people picked up the app and contributed to the data that was collected. Curio was also used to support many great projects and initiatives. We have endeavoured to mention many of those projects below along with the open data tree inventories we uploaded onto the platform.

Collaborative projects supported by Curio

• Morton Arboretum - Chicago Regional Tree Initiative

• Dublin City Council’s Parks, Biodiversity and Landscape Services & School of Geography at University College Dublin - Tree Mapping Dublin

• Sacramento Tree Foundation - Save the Elms Program

• Cambridge City Council - Cambridge City Canopy Programme

• Municipality of Oslo Agency for Urban Environment - Inventory and ecosystem services report hosting

• Friends of Brunswick Park

• Exeter Trees

• Wembley Park Limited

• Washington Square Park Eco Projects

• Coláiste Bríde Enniscorthy

• Enniscorthy Vocational College

• Mountshannon Arboretum - Forester Bernard Carey initiated the Mountshannon i-Tree project, in conjunction with UCD and UK-based consultancy Treeconomics.

• Sidmouth Arboretum

• East Devon District Council

• SLU - Alnarp - Skåne Tree Inventory and support for and involvement in the New Commons and Curio Canopy projects

• Malmö Stad - Malmö Tree Inventory and support for and involvement in the New Commons and Curio Canopy projects

• Göteborgs Stad -

• Halmstad

• Hvilan

• Familjebostader

Open Data Sources Attribution

• The Greater London Authority Datastore - Local Authority Maintained Trees

• NYC OpenData - 2015 Street Tree Census - Tree Data

• Open Data BDN - Street trees of the city of Barcelona

• Open Data Bristol - Trees

• Open Data NI - Belfast City Trees

• Denver Open data - Tree Inventory

• Open Data DK - City of Copenhagen Trees

• Palo Alto Open Data - Palo Alto Trees

• Fingal County Council Open Data - Fingal County Council Trees

• Data SA - City of Adelaide Street Trees

• Open Data Boulder Colorado - Tree Inventory Open Data

• Biodiversity Ireland - Hertitage Trees Ireland

• Birmingham City Council Trees

Uses

The data is free to be used for research purposes subject to the cc-by-nc-4.0 licence and suitable attribution, please see the citation section below

Some potential uses might include;

• Investigations into urban tree biodiversity.
• The development of algorithms for extracting tree attributes via photos or streetview imagery.
• A tree species detection app.
• The detection trees of via satellite imagery.
• Species identfiication via hyperspectral tree.

It worth noting that for most use-cases cleaning, analysis and processing of data will be necessary. The completeness of tree inventory data varies greatly and users were not directed in anyway in terms of how to frame the photos they took and uploaded via the Curio app.

Dataset Structure

TaggedTrees

Number of data points: 2,593,139

The details of an individual tree including its location, species, diameter at breast height (dbh), vitality etc. when available

Images

Number of data points: 27,288

The details of images that were uploaded to the platform. The path to the actual image uploaded, this can be found in uploads directory. The details of what the image was attached to which usually was a ‘Story” that was then attached to a tree are also included.

Uploads:

The set of images referenced in the images data file. The set of images was quite large even when zipped and so was broken up into 10gb chunks. Download each of the chunks and then run unzip on the uploads.zip file

A folder containing downsized versions of the images based on a fixed width has also been included - resized-uploads-width1200.zip

Stories:

The details of a story that was attached to tree

Notes:

The text included in a story/note about a tree.

Conversations & Comments:

Comments grouped by conversations linked to a particular Story

TreeSpecies

The tree species dictionary we built to support the platform. Each TaggedTree has a tree_species_id that references an entry in this dictionary when populated.

TreeSpeciesAliases

The local names across multiple languages that can used to describe a species of tree contained in the TreeSpecies dictionary

Tags and Taggings

Trees could be tagged with details such as diseased, monitored, newly_planted, apples, overhead cables etc. Anything at all really that could later be used to filter, group or identify trees of interest as well describe their state.

Dataset Creation

Curation Rationale

The goal of the Curio platform was to educate, engage and democratised access to environmenatal information. Making the data collected on the platform available in this form is seen as an extension of that mission.

Data Collection and Processing

All data was collected via the Curio app by its community. Where inventory data was uploaded in bulk we preprocessed the data to ensure details such as species information where mapped to the species dictionary we deinfed and that has been included in this release.

Before making the data available on this platform we decided to run face detection and blur any obvious, detectable faces found in the images that have been included.

Citation [optional]

@misc{CurioTreeData, title = {The Curio Tree Dataset}, author = {Conor Nugent and Paul Hickey}, year = {2023}, publisher = {HuggingFace}, journal = {HuggingFace repository}, howpublished = {\url{https://https://huggingface.co/datasets/BreadboardLabs/CurioTreeData}}, }

Dataset Card Authors

Conor Nugent and Paul Hickey

Dataset Card Contact

Conor Nugent