Vlog #5: Keyonte & Devs on Stage, Day 2 (Tableau Conference 2016) #data16
They were knocking out feature after feature until I tweeted that it was getting silly.
- Tableau shared a three-year product vision rather than just next-year features, giving a clearer sense of the company's research direction
- Project Maestro previews visual data cleaning and preparation in a familiar Tableau-style interface
- Hyper, an acquired university research project, promises real-time ingestion and analysis of billions of rows without cumbersome overnight extracts
- Devs on Stage demoed design refinements like distribute evenly, kerning and padding, plus a crowd-pleasing PDF connector and offline mobile workbooks
- Metrics hint at a tiered analytics model: summary metrics, mid-tier discovery visualisations, and detailed granular data
This is a personal vlog from a Tableau conference recapping a keynote and product-reveal session, and its real value is a plain-English digest of a wave of future product announcements and what they signal about the product's direction, rather than a technique to apply.
Recorded on day two of a Tableau conference, covering a morning keynote, a hands-on server training session, and an evening "Devs on Stage" product keynote where a string of upcoming features were demoed back to back.
- A three-year vision, not just next year 4:24
Rather than only previewing near-term features, the team laid out a three-year product direction, which the presenter found reassuring because it shows the underlying research thinking rather than a one-year feature list.
- Visual data prep previewed 5:02
A prototype (Project Maestro) showed cleaning and preparing data visually in an interface similar to existing analytics tools, aimed at letting people see and fix data problems in the same familiar way they build visualisations.
- Faster ingestion for very large datasets 6:02
An acquired research project (Hyper) demoed real-time ingestion and analysis of datasets in the billions of rows, targeting the common pain of overnight extract refreshes and unwieldy large extracts.
- Small design controls with big impact 7:25
New fine-grained design controls — distribute evenly, kerning, padding and margins — make details that were previously fiddly or manual much easier, which the presenter argues matters because careful design visibly changes how well a visualisation communicates its story.
- Getting data out of unusual sources 8:20
A connector able to pull data directly out of PDFs drew the strongest audience reaction of the day, illustrating how removing a data-access barrier can be as exciting as any analytical feature.
- Offline mobile access 8:55
Upcoming offline workbook support on mobile addresses a real limitation: mobile views normally need a live connection, which fails in poor-reception situations.
- Tiered analytics: metrics, discovery, detail 9:21
Metrics hint at a layered model of analytics — high-level summary metrics, a middle layer for exploratory discovery visualisations, and a bottom layer of granular detail — reflecting the view that people shouldn't be expected to land straight in detailed data without a way in.
- Expanding APIs and new chart types 10:00
Continued investment in APIs and client libraries suggests broader programmatic access across the product over time, alongside new chart types chosen from proven, academically grounded visualisation techniques plus added analytical capability such as time series analysis.
- Most of what was shown had no confirmed launch date — some flagged for the next year, some further out — so treat it as direction of travel, not a roadmap you can plan against.
- A feature (custom tooltips/charts) was hinted at again this year having also been shown the previous year, a reminder that repeated previews don't guarantee near-term release.
Watch this if you want a fast, opinionated recap of what a vendor is signalling about its future direction — useful for gauging whether to wait for upcoming capability rather than solve a problem with current tools today.
How this Rollup was made provenance & method
A Rollup is drafted by AI from the video's transcript, then reviewed and edited by Tim. Everything used to produce this one is listed below — the model, the exact prompt, and the source video — so the process is transparent and reproducible.
- Transcription
- On-device — NVIDIA Parakeet v3 for recent videos, OpenAI Whisper large-v3 for earlier ones. The transcript never leaves the machine or gets published.
- Drafting
- Claude Sonnet 5 in the cloud, from that transcript.
- Prompt
- The exact Rollup prompt (v2) — the full system prompt, unedited.
- Source video
- Watch on YouTube
- Drafted
- 5 July 2026 at 09:38
- Reviewed & edited
- 5 July 2026 at 09:41 · by Tim Ngwena
Model + prompt + video is everything you'd need to recreate a Rollup like this yourself. The one thing we don't share is the transcript.
Rights. The video and its transcript are the property of TN Media Ltd. Unauthorised use or download is prohibited. © TN Media Ltd.