- The problem
- Journalists and researchers need to extract information from Arabic texts and compare documents rather than read each one in isolation.
- Our approach
- Analyse Arabic text for entities, keywords, dates, sentiment and recurring phrases, and connect documents through shared words and entities.
- What you can use
- Analyse pasted text and PDF, DOCX and TXT files, extract Arabic text from images with OCR, and explore relationships between documents. PDF and CSV export options are available depending on the plan.
Work & projects
From a newsroom problem to something people can use.
A closer look at our published tools, references and datasets: the problem each addresses, the approach behind it and what is available to use today.
- The problem
- Understanding coverage means tracing what was published, by which source and when, and how narratives were framed over time.
- Our approach
- Archived coverage search with source and timing information, tone analysis and recurrent phrases to explore patterns in discourse.
- What you can use
- Searchable archived coverage, source and timing details, tone and recurrent phrases, a bilingual Arabic–English editorial brief and export options. These are product capabilities, not claimed client outcomes.
- The problem
- Tools for collecting, cleaning, analysing and visualising data are scattered across many sources. Finding an appropriate tool takes time.
- Our approach
- A curated directory organises tools around newsroom tasks, from table extraction to mapping and network analysis.
- What you can use
- More than 120 tools in 10 categories, including Tabula, OpenRefine, QGIS and Gephi, with direct links to the tools.
- The problem
- Working journalists need quick access to commands and concepts they have learned, without restarting a full course.
- Our approach
- Interactive references cover Excel, statistics, Python, Tableau and Power BI, connecting learning with day-to-day practice.
- What you can use
- An Excel reference with 31 commands and functions, a bilingual guide to 26 statistical terms, and further coding and visualisation references.
- The problem
- Research into media framing needs an organised sample that researchers can inspect and reuse.
- Our approach
- A keyword-stratified sample of Gaza war news headlines, derived from a dataset published on Harvard Dataverse.
- What you can use
- A sample of 3,000 headlines across five keyword categories, published on Kaggle for natural language processing and coverage analysis.
Bespoke services / not client case studies
Your brief may call for something different.
The examples above are published products and resources. Media monitoring and discourse analysis, newsroom AI localization and governance, and tools and chatbots connected to your data are bespoke services we scope with you—not client delivery claims attached to these examples.
Explore services and scope