Looker Studio Alternative for Local CSV, Excel, and Parquet Files

Why Looker Studio struggles with local CSV, Excel, and Parquet files

Data Studio (renamed from Looker Studio on April 11, 2026, and formerly Google Data Studio) was designed as a cloud reporting layer on top of Google Analytics, Google Ads, and BigQuery. When you try to use it as a desktop-style data exploration tool on local files, four friction points show up:

1. Strict file upload size and storage caps. The native File Upload connector rejects CSV files larger than 100 MB per file and limits total uploaded CSV storage to 2 GB per Google user (up to 1,000 datasets per user).
2. No native .parquet or .twbx support, and restricted Excel uploads. The CSV File Upload connector only accepts UTF-8 tabular text files (and rejects files with line breaks inside fields, even if escaped with quotes). While Data Studio has a Microsoft Excel connector, it connects to only one worksheet at a time from an uploaded .xlsx file up to 100 MB (capped at 2 GB total Excel storage on the free tier or 5 GB on Pro, with no pivot tables or macros), or you have to convert spreadsheets to Google Sheets (which caps workbooks at 20 million cells). Analyzing a .parquet file requires loading it into a paid BigQuery project first.
3. Every file is uploaded to Google Cloud. You cannot inspect a confidential CSV locally without transmitting the rows to Google's servers.
4. Weak multi-level aggregation grammar. Data Studio lacks Level of Detail expressions (FIXED, INCLUDE, EXCLUDE). Building cohort retention tables or customer-level lifetime value requires clumsy self-blends (capped at 5 tables per blend) or pre-aggregating in SQL.

How to explore CSV, Excel, and Parquet files locally in xmetric

1. Load files directly in the browser. Import a .csv, .xlsx, or .parquet file, or open a .twbx or .twb workbook, in the xmetric app (/app/). Nothing is uploaded to Google Cloud Storage or anywhere else unless you save online.
2. Explore with shelves and marks. Drag dimensions and measures to Columns, Rows, and the Marks card (Color, Size, Detail), and change date levels (YEAR, QUARTER, MONTH, WEEK, DAY).
3. Write LOD expressions instead of fragile 5-table blends. Compute customer-level or category-level denominators inside a single calculated field. To test a formula alone and see the equivalent SQL, use the formula playground (/play/).

Replacing Looker Studio self-blends with a single FIXED LOD

In Data Studio (Looker Studio), dividing a row's metric by a customer's first purchase month or a parent category total requires creating a blended data source joining the table to itself. In xmetric, it takes one calculated field.

Customer cohort retention month index

xmetric calculated field (Months Since First Order):

DATEDIFF('month', { FIXED [Customer ID] : MIN([Order Date]) }, [Order Date])

Equivalent SQL query:

SELECT
  customer_id,
  order_date,
  DATE_DIFF('month', MIN(order_date) OVER (PARTITION BY customer_id), order_date) AS months_since_first_order
FROM orders;

Percent of parent region revenue (EXCLUDE LOD)

xmetric calculated field:

SUM([Revenue]) / SUM({ EXCLUDE [City] : SUM([Revenue]) })

Equivalent SQL query:

SELECT
  region,
  city,
  SUM(revenue) / SUM(SUM(revenue)) OVER (PARTITION BY region) AS share_of_region_revenue
FROM sales
GROUP BY region, city;

Current scope and limits

• Local files supported today: The xmetric app imports .csv, .xlsx, and .parquet files and opens .twbx and .twb workbooks, with shelves, Show Me charts, calculated fields, LODs, and table calcs. The workbook inspector also reads .pbix and .pbip files.
• File size limit: Bound by your computer's available browser tab memory rather than a 100 MB server upload gate.
• Live cloud databases: Direct connections to Postgres and BigQuery are coming.
• Agent commands: Every action in the app is a command. An MCP server, ChatGPT app, and Claude connector are in development.

Spec comparison: xmetric vs Looker Studio (Data Studio) for file analytics