Newsworthy source local news from big data. In 2021 we did a big push to bring data-driven climate journalism to local newsrooms. How will climate change affect local communities? And are these communities transitioning fast enough?
The coverage includes stories on car electrification, changing air travel behavior during Covid-19, industrial greenhouse gas emissions, carbon budget, snow forecasts, heat waves, and much more. These stories are broken down to municipality level and presented in hundreds of local editions, semi-automatically created from carefully curated data.
With automation we are able to bring high-end data journalism to newsrooms with scarce resources.
Newsworthy’s primary audiences are local journalists, local public officials, and politicians. Our reporting reaches people with the power to change local policies, and set local agendas.
Climate reporting is often data-heavy, and might feel daunting for many news reporters. That makes it hard for many smaller newsrooms to get started. Newsworthy makes it possible even for low-resourced publications to do locally relevant climate reporting, by finding relevant angles to the data for each of Sweden’s 21 counties and 290 municipalities.
During 2021, we have had on average around 100 mentions/month from local and national media outlets, who are either using our stories as a base for going even more in-depth, or are simply republishing our articles (source: Swedish media archive Retriever). We have also noticed some un-credited republishing, but these are naturally harder to measure.
To increase the outreach and impact we collaborate with established national outlets – broadcasters, daily newspapers and magazines. The partner newsroom typically publishes a national version of the story, while we publish the local versions on our website.
Description of portfolio:
Newsworthy uses statistical modelling and text automation to make this possible, with statisticians, developers and journalists in the team.
Data is collected and analysed in Python and R (with R doing the heavy lifting in more complex statistical analyses).
Texts are generated by journalists, but assisted by robots. We are automating not to replace human writers, but to empower them. We use a custom-built text templating toolset, making use of grammar rules from CLDR, dictionary data from the university of Gothenburg, and in-house language-technological databases.
Three lessons learned:
1. To automate the right things. We are working in a space between traditional journalism, and the automated text robots that have become popular over the past decade. We involve humans to define newsworthiness, to add interviews for context to data and in-depth analysis. And robots enable us to create hundreds of articles from a small set of human-created templates.
2. To use data to connect the global with the hyperlocal. How does the record-breaking heat wave in your town last week fit into the bigger picture of climate change? How much more common can we expect such heat waves to become? How will it most likely affect that town?
3. Data journalists should start monitoring climate in the same way we are monitoring covid cases, vaccinations and deaths. We need countinous eyes on the key cliamte numbers. Are we transitioning fast enough? Where are we falling behind?
Manifesto for a data-driven climate journalism (Swedish, but works well through Google Translate)
Our climate articles from the city of Umeå (as an example of one of our 300+ local news feeds): https://www.newsworthy.se/nyheter-om/ume%C3%A5-kommun?%C3%A4mne=klimat
Example on partner publication – how the Swedish climate target from transportation are incompatible with a climate budget based on the Paris agreement: https://www.di.se/hallbart-naringsliv/trots-egna-klimatmalet-transportsektorn-spracker-svensk-koldioxidbudget/
Another parnershop publication: Carbon budgets region by region: https://www.sverigesnatur.org/aktuellt/sveriges-koldioxidbudget-racker-fem-ar-hur-ser-det-ut-i-din-kommun/
Team presentation: https://www.newsworthy.se/about-us
Persons nominated: Jens Finnäs, Martin Olsson, Clara Guibourg, Leonard Wallentin