Talk Script - From Field Recordings to Findings (10 min) | AI in Academia - Dataspheres AI

From Field Recordings to Findings - the 10-Minute Talk Script Read-aloud pace: 130 words per minute - the ASHA benchmark for formal speech and the recommen...

From Field Recordings to Findings - the 10-Minute Talk Script Read-aloud pace: 130 words per minute - the ASHA benchmark for formal speech and the recommended pace for data-dense talks. Scripted: 882 words (about 6:50) + a 2:00 live demo + about a minute of pauses and transitions = 10:00. Italic bracketed lines are stage directions - do not read them aloud. Calibration: if you are starting Slide 8 at 7:30, you are on time. Slide 1 - Title · 0:00-0:30 · 63 words Good afternoon. We're Dataspheres AI, and for the next ten minutes we want to talk about something every qualitative researcher in this room lives with - and show you a way out of it. The talk is called From Field Recordings to Findings , because that's the whole promise: one platform that carries a study from the first recording to the final published finding. [ADVANCE] Slide 2 - Industry Pain Points · 0:30-1:30 · 106 words First, an inventory. Tell me if this sounds familiar. [beat] Your interviews are sitting in Zoom recordings, in Google Drive, in audio files on somebody's laptop. Your transcripts live in Otter - which is another subscription you're managing. Your codes live in NVivo - on the one lab machine that has the license. Consent forms and surveys are buried in Qualtrics. The best quote you collected all year is somewhere in a Word document - you'll find it eventually. [pause] And the findings deck is due Friday. [wait for the nods] This fragmentation is not a personal failing - it is the default condition of qualitative research. And it has a price. [ADVANCE] Slide 3 - Fragmentation Has a Price · 1:30-2:30 · 112 words Four prices, actually. Traceability : someone asks, "which voice does this claim come from?" - and answering means an archaeology dig across five tools. Time : hours spent re-finding things you already found. Trust - this one is the heaviest: a reviewer's question, or an IRB audit, that you cannot walk back to the data. Every claim you can't trace is a claim someone else gets to question. And money : a QDA seat can run eight hundred to fourteen hundred dollars a year, transcription is a separate subscription - multiplied by every artifact and every analysis, per person, every year. [beat] None of these costs make the research better. They are the overhead of fragmentation. [ADVANCE] Slide 4 - One Place + the Sample Link · 2:30-3:10 · 85 words So - what if the whole study lived in one place? One datasphere holds the study: recordings, transcripts, codes, datasets, surveys, reports - with dashboards that stay in sync in real time. And we packaged that entire arc into a free, self-paced course - six modules. But rather than tell you about it, let me just show you. This is the sample study behind the course - a real, public research workspace called AI in Academia . Let's step out of the slides for two minutes. LIVE DEMO - 3:10 to 5:10 (2 minutes; cues only, speak freely) 1. Click the sample link on the slide - lands on the AI in Academia datasphere. 2. Corpus guide (~20s): "42 sources on AI in academia - every URL verified live. Critical voices, enthusiastic voices, formal debates." 3. Library, open a transcript (~30s): "Fourteen recordings transcribed - every paragraph timestamped. This one is the Princeton honor-code analysis." 4. Coded Segments dataset (~30s): "Fifty-three verbatim quotes - each with speaker, timestamp, theme, and stance. Click any row, open the transcript, check the coding yourself." 5. Findings page (~30s): "Seven findings - every claim anchored to a quote and a timestamp. And the intercoder reliability page shows our receipts." 6. Return to the deck. Transition line: "Everything you just saw was built with exactly the workflow the course teaches. Here are its six modules." Running long? Skip the Findings stop - Module 6's slide covers it. The next six slides are 30-45 seconds each; the exit line for each is on the slide in gold. Slide 5 - Module 1: Your Research Workspace · 5:10-5:40 · 62 words Module one: your research workspace. A datasphere is a home for the whole study - and you hold the keys. Roles and access controls scoped to your study. Privacy walls around participant spaces. Audit trails on every AI-assisted step. That's the IRB conversation, answered on day one. And the workspace you just saw is public - anyone here can inspect it tonight. [ADVANCE] Slide 6 - Module 2: Collecting Data · 5:40-6:10 · 59 words Module two: collecting data. Everything lands in one place - interviews, video, documents, images, participant journals, and continuous surveys, uploaded from any device, straight into the study workspace. For the sample study, that meant forty-two sources - and every single URL was verified live before it entered the corpus. Intake with a paper trail, from the very first file. [ADVANCE] Slide 7 - Module 3: Transcription · 6:10-6:45 · 69 words Module three: transcription for real fieldwork. Voi