What is it worth, and how sure are you?
A valuation workshop that shows its reasoning: where each number came from, which assumptions carry the result, and what the market price already assumes.
Collect
Run the Python collector in Google Colab. It retrieves available annual filings from SEC EDGAR, prices and statements from Yahoo Finance, and macro series from FRED, then scores the data it found.
Import
The collector writes one file, TICKER_dashboard.json. Import it here. Everything after that runs in your browser; nothing is uploaded anywhere.
Interrogate
Read the briefing, move the assumptions, and see which ones matter. Snapshots from the previous studio version still open.
Briefing
Ranked by how much each point should change your view. Every statement is computed from this snapshot.
How the value is built.
From cash flows to a share price
Discount rate
Forecast against the record
Solid is reported, tinted is the base forecast. A forecast that breaks sharply from the record needs a reason. The last view shows how much of today’s value each future year contributes.
Year-by-year forecast
Cross-checks
Independent reference points. None of them feed the model.
Change an assumption. Watch the value.
The same model, recalculated as you move. Start from a preset, or find the point where the model agrees with the market.
Two-way sensitivity
Green cells are above today’s price, red below. The outlined cell is your current setting.
How uncertain is it?
Scenario cloud
Every scenario placed by two of its inputs, with height showing the value it produced. Drag to rotate. Planes mark P10, median, P90 and the market price.
Shape of the distribution
Moments and tails of the simulated values.
Chance the value exceeds
Paths, not just endpoints
Each scenario implies a ten-year path. Bands show where half and four-fifths of the paths run; thin lines are individual scenarios.
Which assumptions carry the range
Share of the spread in simulated values explained by each input (squared rank correlation, normalised).
One assumption at a time
Each dot is a scenario. The line is the median value across that slice of the input.
Driver table
For each input: what the model assumes, how uncertain it is, what one standard deviation does to value on its own, and the level at which the model would agree with today’s price.
Simulation settings and convergence
Running mean with its 95% numerical error band. The band shrinks with more scenarios; it says nothing about whether the assumptions are right.
The record behind the assumptions.
Revenue growth, year by year
Operating margin
Growth quality
How fast, how steady, and whether it is slowing.
Profitability and cash
Whether reported profit turns into cash, and what it costs to grow.
Operating fundamentals
Measured from the statements. These describe the business; they are not forecasts.
How the stock behaves.
Risk measures
Descriptive statistics from daily returns. They characterise the past year, not the next one.
What it moves with
The economy around it.
Official FRED series chosen for this sector. Associations with the stock are historical and descriptive; none of them alter the valuation.
What to check in the filings
Company-level drivers that public macro data cannot measure.
Break it down.
Click any block to open it. The box refills with what that number is made of.
Regulatory capital.
Each figure is labelled by legal entity. A subsidiary’s Call Report is not the parent’s consolidated ratio.
Can you trust the inputs?
Every check is explicit. A low score does not make the valuation wrong; it tells you how much of it rests on assumption instead of evidence.
Model cautions
Raised by the collector while building this snapshot.
Where each input came from
Flagged rows were assumed or reconstructed, not observed.
The numbers, in full.
Everything shown above, exportable.