Ask it anything.
Report it anywhere.

Power BI, Excel and Tableau connect to Data Oil through the PostgreSQL driver they already ship with, and a question typed as a sentence is answered with the rows behind it. Below is what each of them gets back from one real table.

700 mobile users
11 columns
5 behaviour classes
Every figure checked against the SQL beside it

Power BI

A report in the tool your analysts already use

Get data, PostgreSQL database, and the tables are in the navigator. The heavier someone uses a phone, the harder it works: battery drain climbs from 455 to 2,701 mAh a day across the five classes.

Power BI DesktopGet data › PostgreSQL database › user_behaviour

700users in the table

5.3 hscreen on, a day

930 MBdata used, a day

1,525 mAhbattery drained, a day

Battery drained a day, mAh, by behaviour class

Screen on a day, hours

Class 11.5 h

Class 23.0 h

Class 35.0 h

Class 46.9 h

Class 510.1 h

The query behind the reportThe statement it sends

SELECT "User Behavior Class" AS class,
       count(*) AS users,
       round(avg("Battery Drain (mAh/day)"), 0) AS battery_mah,
       round(avg("Screen On Time (hours/day)"), 1) AS screen_hours
FROM user_behaviour_dataset_rows
GROUP BY "User Behavior Class"
ORDER BY class;

Class	Users	Battery, mAh	Screen, hours
1	136	455	1.5
2	146	884	3.0
3	143	1,515	5.0
4	139	2,106	6.9
5	136	2,701	10.1

Excel

A pivot table on live data, not last month's export

From Database, From PostgreSQL Database — the same connection, the same driver. Here the pivot shows who owns what: every phone carries every kind of user, and no cell holds fewer than 21 or more than 33.

ExcelData › Get Data › From PostgreSQL Database › PivotTable

	A	B	C	D	E	F	G
4	Users by device	Class 1	Class 2	Class 3	Class 4	Class 5	Total
5	Google Pixel 5	32	26	33	24	27	142
6	OnePlus 9	25	31	25	26	26	133
7	Samsung Galaxy S21	26	30	21	33	23	133
8	Xiaomi Mi 11	31	27	32	27	29	146
9	iPhone 12	22	32	32	29	31	146
10	Grand total	136	146	143	139	136	700

The query behind the pivotThe statement it sends

SELECT "Device Model", "User Behavior Class", count(*) AS users
FROM user_behaviour_dataset_rows
GROUP BY "Device Model", "User Behavior Class"
ORDER BY "Device Model", "User Behavior Class";

Tableau

Every user on one chart

All 700 rows, one dot each. App time and battery drain rise together, and the five classes sit in bands along that line that never overlap — each one begins where the one below it ends. Class 1 averages 60 minutes in apps, class 5 averages 541.

TableauConnect › PostgreSQL › user_behaviour_dataset_rows

Class 1 · lightestClass 2Class 3Class 4Class 5 · heaviest

The rows behind the chartThe statement it sends

SELECT "User ID", "App Usage Time (min/day)", "Battery Drain (mAh/day)",
       "User Behavior Class"
FROM user_behaviour_dataset_rows;

Ask it in a sentence

No SQL, no report to build. Just the question.

The answer names the rows it rests on, and each one here is checked against the SQL shown with it — so you can run the check yourself, in any of the tools above.

You askedDo Android or iOS users drain more battery?

Data Oil

iOS, by a little. The median iPhone user drains 1,538.5 mAh a day against 1,489 for Android, and the averages agree: 1,589.5 against 1,508.2. There are 146 iOS users and 554 on Android, so the Android figure rests on nearly four times as many people.

Evidenceuser_behaviour_dataset_rows · Grouped by operating system · all 700 rows

The SQL that checks itRun it yourself

SELECT "Operating System", count(*) AS users,
       median("Battery Drain (mAh/day)") AS median_mah,
       round(avg("Battery Drain (mAh/day)"), 1) AS average_mah
FROM user_behaviour_dataset_rows
GROUP BY "Operating System";

You askedWhat separates the heaviest users from the lightest?

Data Oil

Everything, and in step. Class 5 users keep the screen on 10.1 hours a day against 1.5 for class 1, spend 541 minutes in apps against 60, have 89 apps installed against 15, use 1,975 MB of data against 202, and drain 2,701 mAh against 455 — from 5.9 to 9.8 times as much, on every measure.

Evidenceuser_behaviour_dataset_rows · Classes 1 and 5 · 136 and 136 users

The SQL that checks itRun it yourself

SELECT "User Behavior Class",
       round(avg("Screen On Time (hours/day)"), 1) AS screen_hours,
       round(avg("App Usage Time (min/day)"), 0) AS app_minutes,
       round(avg("Number of Apps Installed"), 0) AS apps,
       round(avg("Data Usage (MB/day)"), 0) AS data_mb,
       round(avg("Battery Drain (mAh/day)"), 0) AS battery_mah
FROM user_behaviour_dataset_rows
WHERE "User Behavior Class" IN (1, 5)
GROUP BY "User Behavior Class";

You askedWhich phone uses the most data?

Data Oil

The iPhone 12, narrowly. Its users average 966 MB a day; the Google Pixel 5 is lowest at 898. Across all five models the spread is under 8%, so the phone says far less about data use than how its owner uses it does.

Evidenceuser_behaviour_dataset_rows · Grouped by device model · 5 models

The SQL that checks itRun it yourself

SELECT "Device Model", round(avg("Data Usage (MB/day)"), 1) AS data_mb
FROM user_behaviour_dataset_rows
GROUP BY "Device Model"
ORDER BY data_mb DESC;

Connecting

One connection, three tools

Power BI Desktop

Get data → PostgreSQL database. The navigator lists your tables and Import loads them — measured on a real deployment.

Excel

Data → Get Data → From Database → From PostgreSQL Database, then a PivotTable over what it returns.

Tableau

Connect → PostgreSQL. The same host, port and database, and the same account.

Each one needs the host and PostgreSQL port of your deployment, the database by name, a Data Oil account, and SSL on. PostgreSQL tools has what was measured, tool by tool, and what is declined.

Ask your own data.

The proof of concept starts with your tables, your questions and the tools your team already uses.

Get started Contact

Ask it anything.Report it anywhere.

A report in the tool your analysts already use

Battery drained a day, mAh, by behaviour class

Screen on a day, hours

A pivot table on live data, not last month's export

Every user on one chart

No SQL, no report to build. Just the question.

One connection, three tools

Ask your own data.

Ask it anything.
Report it anywhere.