Guide

Ask the Data — Full Guide

How to set up your session, unlock the AI chat, and ask your streaming data the right questions.

Prerequisite: Upload your DSP data first

Ask the Data requires an active data file in the DSP Dashboard. If you haven't uploaded your CSV/JSON export yet, follow the DSP Dashboard guide to upload your file before continuing.

What Is Ask the Data?

Ask the Data is an AI chat interface that has direct access to your uploaded streaming dataset. Instead of manually filtering and sorting your spreadsheet, you ask questions in plain English and get structured answers — complete with specific numbers from your data. It can identify trends, compare platforms, rank tracks, flag geographic anomalies, and explain patterns that would take hours to surface manually.

Your data, not general knowledge

Every answer is grounded exclusively in the data you uploaded. The AI does not make up numbers or supplement with external data. If something isn't in your file, it will tell you.

3 credits per session, 10 turns

Each Ask the Data session costs 3 credits and gives you 10 turns. Credits are deducted when you send your first message and confirm the date range.

Setting and Locking Your Date Range

Before Ask the Data activates, you must define a date range. This range determines which rows from your uploaded data the AI can access during the session. Once you confirm the range and unlock the session, it cannot be changed.

1

Click "Ask the Data" in the DSP Dashboard

With your data uploaded, click the "Ask the Data" button in the dashboard header or sidebar. The session setup panel appears.

2

Choose your date range

Use the date pickers to set a start and end date. The panel shows how many rows fall within your selected range, so you can confirm you're capturing the right data before spending credits. Presets for "Last 30 days", "Last 90 days", and "All data" are available.

Tip: A focused date range (e.g., a specific quarter or release window) produces more precise answers than asking the AI to analyze years of data in 10 turns.
3

Confirm and unlock the session (3 credits)

Review the date range and row count summary, then click "Unlock Session." A confirmation dialog shows the credit cost. Confirm to proceed. Your date range is now locked and the chat interface activates.

Important: The date range locks permanently for this session. If you need to analyze a different time window, you will need to start a new session (3 more credits). Choose carefully.

8 Questions That Get Great Answers

Ask the Data performs best when questions are specific and reference actual dimensions in your dataset (platform, country, track, date range). These examples consistently produce high-value answers.

Which of my tracks had the highest stream-to-revenue ratio in the selected period, and what platform drove that?
Show me my top 5 countries by streams and compare their revenue-per-stream rates. Which market is the most monetarily efficient?
Which platform showed the most growth in monthly streams between the first and second half of my date range?
Identify any tracks that had unusually high streams in a specific country that doesn't appear in my top 10 markets overall. Are there geographic outliers I should pay attention to?
What percentage of my total revenue came from Spotify versus non-Spotify platforms? Has that ratio shifted over the course of the date range?
Which of my releases from the last 12 months has the highest stream velocity (streams per day since release)? Is it still accelerating or declining?
Compare my top 3 tracks' streaming patterns by day of week. Do any tracks consistently over-perform on specific days, suggesting playlist rotation patterns?
What is my estimated monthly streaming income trend over the selected period? Is my catalog's revenue baseline rising, flat, or declining between releases?
Follow-up strategy: After getting an answer, use a follow-up turn to go deeper: "Break that down by month" or "Which specific tracks are driving that?" Follow-ups don't need to repeat context — the AI remembers the full conversation.

What Ask the Data Can and Can't Tell You

What it CAN do

  • Sum, rank, compare, and calculate ratios from your data
  • Identify trends over time within your date range
  • Flag geographic and platform outliers in your dataset
  • Calculate derived metrics (revenue per stream, growth rates, etc.)
  • Identify your fastest-growing tracks by velocity
  • Compare performance before and after a specific date (e.g., a release)

What it CANNOT do

  • Access data outside your uploaded file or selected date range
  • Tell you why a trend happened (it can describe it, not explain algorithmic cause)
  • Access real-time or live streaming data
  • Compare your numbers to industry benchmarks (use Metrics Deep Dive for that)
  • Analyze columns not present in your uploaded file
  • Guarantee revenue projections — estimates are based on your historical data only

Managing Your 10-Turn Session

Each Ask the Data session gives you 10 turns. Here's how to get the most out of them.

Plan your questions before unlocking

Write out 5–6 questions you want answered before you spend the credits. Starting with a clear agenda prevents wasting turns on discovery questions that could have been answered by looking at the dashboard charts.

Use turns for analysis, not retrieval

Don't waste turns asking for raw totals you can read directly from the dashboard (e.g., "What are my total streams?"). Use turns for questions that require calculation, comparison, or pattern analysis across your dataset.

Chain follow-ups strategically

Start broad (e.g., "Which platform is performing best?"), then follow up with specifics ("Break that down by month" → "Which tracks drove that result in July?"). Three well-chained turns can produce more insight than six independent questions.

End with a summary request

Use one of your final turns to ask: "Based on everything we've discussed, what are the 3 most actionable takeaways from my data?" This synthesizes the session into a usable action list.