Home / Services / Practice 03
Data analytics and visualisation
Text and audience intelligence at national scale, and decision dashboards for people who have to act on them.
Practice areas
- Big data and text analytics — mining large Arabic and English corpora for themes, entities and change over time.
- Social media analytics and strategy — audience intelligence, content and sentiment analysis, and the strategy that follows from it.
- Web analytics — measurement design and reporting for public-facing digital services.
- Financial market analytics — as an authorised MetaStock education partner we deliver training on MetaStock Pro, Xenith and associated add-ons.
- Visualisation and dashboards — interactive dashboards for executives and boards.
Forrester Research cited our work in Arabic natural language processing, big data and text analytics in its November 2015 report The Gulf Cooperative Council’s Big Data Opportunity: How the GCC Can Use Big Data To Be More Competitive.
The Arabic layer
Standard analytics platforms treat Arabic as a character set rather than a language. We extend them: our Social Intelligence Analyzer runs as an add-on to the Pulsar social media monitoring system, adding content and sentiment analysis for Modern Standard Arabic and Kuwaiti dialect. It has been used by the Kuwaiti Parliament and Kuwait TV to analyse public trends, and was published as a Pulsar case study in September 2016.
Analytics through a conversation
When the analyst is an AI agent
Two of the systems we run at Kuwait International Law School are analytics platforms whose interface happens to be natural language. They are worth describing here because they show what institutional analytics looks like when the reporting bottleneck is removed.
The Academic Intelligence Assistant lets staff ask questions of live student records in Arabic or English and get back pass and fail rates by course and term, demographic breakdowns, and cohort comparisons — with genuine statistics behind them: Pearson correlation, R-squared, p-values, linear regression and ANOVA, plotted as scatter and trend charts in the chat itself. A risk-scoring engine rates each student 0–100 and sorts them into tiers, so advisors can see who needs attention while intervention is still possible.
The evaluation analysis platform turns each term’s course and faculty surveys into a report management can act on. Positivity is scored per question rather than averaged, making instructors and questions directly comparable, and the full set of free-text comments is read rather than sampled — with a 10% threshold before a theme is reported, and serious matters surfaced regardless of frequency.
Both read live institutional data, strip identifiers before anything leaves the database, and run inside the client’s own network. Full descriptions are on the generative AI page.
Start here
Tell us what your text is doing wrong.
Send us a sample — a set of citizen comments, a document archive, a support inbox — and we will come back with a written read on what is achievable, what it would take, and whether you need us at all.
Send a sample of your text with the form. We reply in writing.