EconData<\/a>. The platform centralises economic and financial data into a single, consistently structured format, ready to plot or model. When a new release lands, ingestion scripts pull it in, validate it against what came before, and push it through chart templates that auto-update. The analyst doesn’t reshape anything. The chart refreshes itself.<\/p>\n\n\n\nThe platform is also built using standards like SDMX \u2014 Statistical Data and Metadata Exchange \u2014 promoted by Eurostat and used by the IMF and the World Bank. That matters for two reasons. It makes the data interoperable with tools analysts actually use: a website view for casual users, an Excel add-in for spreadsheet people, and an API for anyone working in R or Python. And it bakes metadata into the data itself, so quality control can be automated rather than done by eye.<\/p>\n\n\n\n
The vintage problem<\/h2>\n\n\n\n
One of Codera’s less obvious contributions is how it handles revisions. Economic data changes after the fact, sometimes heavily. A model trained on today’s cleaned-up historical series will look accurate in backtesting and underperform in practice, because it never had to work with the provisional numbers analysts actually saw at the time.<\/p>\n\n\n\n
Codera stores every vintage: what was published on the day, not just what it was later corrected to. That makes it possible to show, for example, how National Treasury’s debt projections have ratcheted upwards budget after budget, each one expecting a stabilisation that never came.<\/p>\n\n\n\n
The same philosophy extends to timeliness. Official house price indices run months behind. Codera is building alternative-data indicators that estimate current-month prices from private sources, and extending the approach across industries \u2014 weekly farm-gate prices feeding into food inflation outlooks, input costs helping anticipate producer prices before Stats SA confirms them.<\/p>\n\n\n\n
Steenkamp was careful here. Real-time private-sector data sounds compelling, but it carries its own hazards: odd seasonality, aggregation traps, and selection bias. “Pure data science can get you into trouble with alternative data,” he said. Statistical training still earns its keep.<\/p>\n\n\n\n
The AI question<\/h2>\n\n\n\n
Asked how artificial intelligence is changing Codera’s work, Steenkamp’s answer was more cautious than the industry average. The team uses AI to automate lower-stakes tasks and for suggestions, but still builds its core systems from scratch. Two reasons: the work has to be correct and fixable when it breaks, and Codera doesn’t want its models training someone else’s.<\/p>\n\n\n\n
His advice to young analysts followed the same logic. Learn to build things from scratch. Understand frameworks deeply. Don’t let AI become a substitute for knowing how the machinery works, because the machinery keeps changing and the people who can adapt are the people who understand it.<\/p>\n\n\n\n
A wish list<\/h2>\n\n\n\n
Pushed on what data he most wants to see, Steenkamp named two gaps.<\/p>\n\n\n\n
The first is public data that is already meant to be public. Many government departments have a legal mandate to publish and don’t, for reasons ranging from funding to capability to unresolved data quality issues. The cost is that researchers, journalists and policymakers can’t monitor performance or diagnose what’s actually broken.<\/p>\n\n\n\n
The second is a framework for responsible data sharing. South Africa has reasonably developed rules for data protection. It has close to nothing on how public institutions should share data with each other, with private firms, or with researchers. Without that, Steenkamp argued, a thriving domestic tech sector is hard to imagine.<\/p>\n\n\n\n
The through-line<\/h2>\n\n\n\n
The argument running through Codera’s charts, its real-time experiments and Steenkamp’s policy critique is a single proposition: data is infrastructure. Published as an afterthought \u2014 locked in PDFs, scattered across incompatible formats, guarded by institutions that aren’t quite sure why \u2014 it becomes something analysts spend their time wrestling rather than using.<\/p>\n","protected":false},"excerpt":{"rendered":"","protected":false},"author":1,"featured_media":91195,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1409],"tags":[],"newsletter-post":[],"site":[],"class_list":["post-91194","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-webinar"],"acf":{"post_style":"bc","show_on_front":"Yes","link_through":"Yes","big_number":"","big_number_caption":"","big_number_link":"","big_number_background":"","big_number_text_colour":"#000000","big_number_icon":false,"big_number_wide":"yes","featured_chart":false,"flourish_chart_id":"","flourish_sub_title":"","flourish_chart_width":"medium","is_newsletter_post":"No","chart_url":"","background_colour":"#0089AA","text_colour":"#FFFFFF"},"_links":{"self":[{"href":"https:\/\/outliereditor.co.za\/index.php\/wp-json\/wp\/v2\/posts\/91194","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/outliereditor.co.za\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/outliereditor.co.za\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/outliereditor.co.za\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/outliereditor.co.za\/index.php\/wp-json\/wp\/v2\/comments?post=91194"}],"version-history":[{"count":3,"href":"https:\/\/outliereditor.co.za\/index.php\/wp-json\/wp\/v2\/posts\/91194\/revisions"}],"predecessor-version":[{"id":91200,"href":"https:\/\/outliereditor.co.za\/index.php\/wp-json\/wp\/v2\/posts\/91194\/revisions\/91200"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/outliereditor.co.za\/index.php\/wp-json\/wp\/v2\/media\/91195"}],"wp:attachment":[{"href":"https:\/\/outliereditor.co.za\/index.php\/wp-json\/wp\/v2\/media?parent=91194"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/outliereditor.co.za\/index.php\/wp-json\/wp\/v2\/categories?post=91194"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/outliereditor.co.za\/index.php\/wp-json\/wp\/v2\/tags?post=91194"},{"taxonomy":"newsletter-post","embeddable":true,"href":"https:\/\/outliereditor.co.za\/index.php\/wp-json\/wp\/v2\/newsletter-post?post=91194"},{"taxonomy":"site","embeddable":true,"href":"https:\/\/outliereditor.co.za\/index.php\/wp-json\/wp\/v2\/site?post=91194"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}