Analytics & Reporting

Ask your master data anything, without exporting it

Report Studio is a full authoring environment built into the hub. Drag fields into wells, pick a visual, add calculations, publish it to a dashboard and let people subscribe to it — over golden records, quality scores, match results, reference data and pipeline runs, on whichever engine your hub runs on. Nothing leaves the database to be analysed.

MDM Studio Design Map Govern Trust
Visuals

Seven ways to show it, one way to build it

Every visual is authored the same way — fields on the left, wells in the middle, properties on the right — so learning one teaches you the rest. Switch a report's visual and the wells it needs come with it.

Record & grid

Tb
TableMany records, with drill-through
Rc
RecordOne record and everything related
Fm
FormA single record, laid out

Aggregate

Mx
Pivot matrixRows × columns × values
Pv
PivotGrouped, with a drill ladder
Ch
ChartTrend and distribution
Sc
ScorecardKPI cards, split by any dimension

Chart types

Column & BarGrouped or stacked
Line & AreaTrend over time
Pie & DonutShare of a whole
ScatterTwo measures, correlated
RadarMany dimensions at once
TreemapNested magnitude
FunnelStage-to-stage fall-off
WaterfallWhat moved the total

Charts can mix marks on one axis — plot one measure as bars and another as a line — and any chart can be split into small multiples, one panel per value of a dimension, on a shared scale.

Authoring

Built by dragging, not by writing SQL

Fields, canvas, properties

Three panes. Drag a field into Rows, Columns, Values, Filters or Small multiples and the preview redraws as you go. Undo and redo the whole session; save with a keystroke.

Filters that explain themselvesNEW

One operator vocabulary across every visual. Relative dates ("last 30 days") resolve as the reader opens the report. A filter can be required, a parameter, or carry a default — and the reader sees which.

Dropdowns where they helpNEW

A filter on a field with few enough distinct values becomes a picker automatically — the threshold is one central setting, adjustable per person or per model. Values are read through the same governed door as the data, so masking and row-level rules still apply.

Calculated fieldsNEW

Write an expression once and use it like any other field. Expressions are compiled to SQL for the hub engine you are on — nothing an author typed is ever sent to the database as text.

Aggregate control

Sum, average, count, distinct count, min and max — with ordering, Top N and an "Other" bucket, having-clauses, and show-as windows: % of total, % of a grouping, running totals and rank. Group a date by day, week, month, quarter or year without adding a column — the truncation is computed by whichever hub engine you are on, and the report echoes the grain it actually applied, because a silently ignored grain mislabels an axis.

Column toolsNEW

One rule set per column, drawn five ways: an indicator, a heat bar, a badge, a range against a target, or a delta against another column. Change a threshold and every mark on that column moves together — a bar and a badge on one column cannot disagree, because there is only one place the answer lives. The same marks reach matrix and pivot measures, and row and column headers.

Rules that read your data, not just your numbersNEW

Thresholds and bands for numbers; contains, starts with, is one of and wildcards for text; on, before and in the last 30 days for dates, resolved as the reader opens the report. A value can name another column, so every domain is judged against its own target rather than one global number. A blank cell matches only is blank — an unmeasured score is never painted as a zero.

Conditional formatting

Colour scales, icons and value rules on table columns, and banded gauge and bullet cards that paint in the direction being scored — so lower-is-better reads correctly. The band editor warns about gaps and overlaps, and can suggest bands from the data's own distribution — with the AI narrating the choice, and the arithmetic deciding it.

Back to the report you were inNEW

The report picker opens with a Recent list — the few reports you actually work in, ranked by how often and how recently you opened them, each showing when you last had it open. It is kept per person and resolved against the reports the server says you may open, so one that was deleted or is no longer yours simply is not on it. The property panes fold as well, and remember what you folded — each closed group states what it is hiding, and a well opens itself the moment you drag a field toward it.

Dashboards

One filter, many reports

Dashboards in MDM Studio are a canvas, not a fixed grid. Tiles reflow, fit, fill or lock their aspect; the canvas takes a colour, image, pattern or grid; and titles are formatted like text, not like settings.

Unified filters across sourcesThe same business idea — a region, a domain, a period — is often a different column name in every report on the page. Bind one filter to the field each source actually uses and it drives them all at once.
It tells you where it does not reachA tile whose source has no matching field is flagged, not quietly left showing unfiltered numbers. You choose per filter whether an unbound tile flags, ignores or blanks — and flagging is the default.
…and lets you propose one yourselfNEWDetection only pairs fields that share a name, which is why a dashboard so often has nothing to unify: customer_id here is cust_id there. Pick the field on each source by hand instead. Suggestions are offered with their evidence — same key, same type, 80% shared values — but nothing is ever bound without you confirming it.
Every tile is grabbableNEWMove a tile from anywhere on it, resize it from any of eight edges, or nudge it with the arrow keys — so a tile scrolled past the top of the canvas, hidden behind another, or taller than the screen is never stranded.
It refuses unsafe matchesFields that merely share a name are not the same field. Where five sources spell "status" five incompatible ways, the detector says so and declines to unify rather than silently filtering the wrong rows out.
Report list by areaThe toolbox groups every report by its area and remembers what you collapsed — and a search always opens the areas that match, so a filtered result is never hidden behind a fold.
One filter concept
Golden recordsDQ exceptionsMatch pairsReference dataPipeline runs
BOUND PER SOURCEone value · many field names
ChartPivotScorecardTable
Sharing

Publish it, save your view, get it emailed

Saved views

Any set of filters, sorts and drill positions can be saved as a named view, made the default, or shared with the report — so "the version I actually look at" is a link, not a ritual.

SubscriptionsNEW

Schedule a report or a whole dashboard to arrive by email. Each run executes as the subscriber, so nobody is ever mailed rows they could not open themselves — and when no mail server is configured the run is recorded as skipped, not as sent.

Export & print

CSV and Excel are generated on the server against the same governed query the screen ran — so an export is the report, not a re-derivation of it. Print keeps the layout.

Pipeline & job notificationsNEW

Subscribe to a job, a job group or an execution package and hear about starts, successes and failures in-app or by email. One incident sends one letter — the most specific scope wins, so watching a job and its group does not double up.

Drill, up and down

Click into a bar, a cell or a KPI card and the report walks the ladder you defined — down a level, up again, or straight to first or last. The pivot follows the same ladder as the chart.

Governed by the same rules

Reports read through the product's own access layer. Sensitivity clearance, row-level access groups and tenant isolation apply to a report exactly as they apply to the record screen, and filter and sort keys are allow-listed against the source catalog.

Honesty

A chart that cannot tell the truth says so

Most reporting tools will draw something no matter what you feed them. A picture of the wrong thing is worse than no picture, because nobody checks it. MDM Studio refuses instead, and names the reason on the visual.

Marks encode their valueA gauge or bullet resolves its scale through a stated ladder — authored, shared, format, target, previous, data — and prints the ends it chose, so you can see what the needle is measured against.
Shares refuse negativesA pie, donut or treemap cannot represent a negative as a slice. Rather than dropping the row silently, the visual declines and says which value it could not draw.
Unknown is not zeroA value that could not be computed renders as unavailable with its reason, never as a confident zero. An empty group keeps its own identity instead of folding onto a dash.
A total is not on the values' scaleNEWShow or hide grand totals, row totals and subtotals as the report needs them. Marks reach total cells only when you ask: a rule reading "≥ 95 healthy" is a claim about a percentage, and the sum beneath it has nothing to do with 95. The trap is named next to the switch rather than left to be discovered.
Subtotals are weightedAn average of averages is not an average. Subtotals over a mean are weighted by the records behind them, and a distinct count refuses to be re-aggregated at all.
Everywhere it shows

The same number, formatted the same way

A display format is governed once on the attribute, in the data dictionary, and every surface honours it — grids, chart data labels, axis ticks, KPI cards and exports.

Formats are display onlyNEWNamed presets over one mask engine, proposed automatically from the source column's type and scale. The raw value is what filters, copies, exports to CSV, compiles into SQL and rides every API payload — formatting never changes the data.
Light and darkNEWThe whole product, charts included, flips from one preference stored on your account and painted before the first pixel. The palette is measured for contrast, not chosen by eye.
Flat, raised or sunkenNEWA report takes a depth, and everything in it follows: cards, headers, rows and chart bars, and now the marks inside the cells — pills, heat bars and heat-map tiles too. Depth is added; nothing is recoloured, so a look can be switched on a finished report without re-checking a single contrast ratio.
An accessible chart paletteEight slots, measured against the background in both themes. Series colour is assigned from that set rather than picked per report, so two reports on one dashboard agree.
What you can report on

Master data, and everything that happened to it

Report Studio reads the product's own governed sources — no warehouse hop, no nightly extract, no second copy to secure. The dashboard's own domain-health and severity queries are published here as sources, carrying the same access scoping they have on the dashboard — so a narrowed steward never sees their slice of records beside a whole domain's exception count.

GR
Golden recordsAny entity, any domain
DQ
Quality & exceptionsScores, rules, remediation
MT
Match & mergePairs, decisions, lineage
RD
Reference dataSets, releases, crosswalks
PL
Pipeline runsJobs, groups, packages
GV
GovernanceRequests, approvals, audit
SR
Source objectsProfiling and staging
SQ
Custom queriesGoverned, allow-listed
DH
Domain healthRecords, score, stewards
SV
Exceptions by severityOpen, closed, by domain
Golden Record Explorer

Follow one record through everything it touches

A report answers a question you already knew to ask. The Explorer is for the other kind — somebody hands you a key and asks whose it is. Start on an entity, walk its relationships tab by tab, and make any related row the new focus with a breadcrumb trail behind you.

Search a key through its relationshipsNEWA Sales Order key finds the Customer that owns it. Searching used to match only the entity you were looking at, so you had to switch entity, read the join column off the row, switch back and filter by that value. The reach is a toggle, and the result names which related entities were searched.
…and it says what it could not searchNEWA relationship with no attribute pair, or one whose entity you have no access to, is named rather than quietly omitted — because a silently narrowed search is indistinguishable from a record that is not there.
Row counts on every relationship tabNEWSee which relationships actually hold rows without clicking each one. A count that could not be computed draws a dash, never a zero — “none related” and “this relationship cannot be resolved” are different answers.
One hierarchy rail, never blendedTiers, record trees, managed hierarchies and value hierarchies, each kept as itself, persisted across the Explorer and Data Quality, and labelled with the business names the model declares.
An empty grid says which kind of emptyNEW“Nothing matched that key”, “nothing in this branch” and “this entity was never published” are three different problems. The grid names the one it means, and never tells you to publish an entity that is already published.

Bring a question you cannot answer today

The fastest way to judge a reporting layer is to point it at a real question. Bring one, and we will build the report with you in the session.

Request a demo All capabilities