Data reference - Adriel Knowledge Base

Introduction

Baidu Analytics (百度统计, Baidu Tongji) is Baidu’s web analytics platform for measuring site traffic, visitor behavior, and promotion performance in the Chinese market. It reports on page views, visits, visitors, engagement, and conversions for a registered site, and — for sites running Baidu promotion — on impressions, clicks, cost, and return. The connector reads from the Baidu Tongji REST API (v1). It authenticates with a username and access token rather than OAuth. Each data source is scoped to a single site and a single report type chosen at creation. The report type maps to a specific Baidu report (for example, Search Term, Visited Pages, or Website Overview) and determines both the metrics returned and the primary breakdown each row represents. When several sites are selected during setup, Adriel creates one data source per site. As an analytics connector, visit, visitor, engagement, and conversion data are standardized so widgets can aggregate this source alongside other connectors. To connect this data source, see How to connect Baidu Analytics to Adriel.

Data refresh strategy

Architecture data

Architecture covers the account’s registered sites. The site list is read live from the Baidu Tongji API at connection and data source creation time (it is not separately cached), and each site’s identifier scopes every subsequent data query. The report type is fixed per data source and cannot be changed without recreating the data source.

Reports data

Reports data uses a daily-cache model. On first connection, historical daily data for the last 180 days is backfilled. After that, a scheduled refresh re-fetches the most recent 30 days on each run, twice per day at 02:00 and 14:00 UTC. Baidu Tongji itself processes data with a short delay (typically a few hours), so very recent rows can still shift between refreshes. Older data outside the 30-day window is not re-fetched, because it rarely changes.

Architecture levels

Account → Site The account is the set of credentials (username and access token). A site is a single registered property, identified by its domain and site ID, and is the reporting boundary for every query. Baidu Analytics has no ad-hierarchy analog; the report type selected at creation, not a drill-down level, determines what each row represents.

Date range limits

Limit	Value
Initial history on first connection	180 days
Rolling refresh window	30 days

Data earlier than 180 days before the first connection is not backfilled and is unavailable. There is no documented per-query maximum range beyond the history available in the cache.

Metrics

How to read the columnsCode style like pv_count is the literal Baidu Tongji API field name returned for the report. Italic text describes a value that Adriel derives rather than reads directly from a single API field.Data type vocabulary: Number, Currency, Percentage, Ratio, Duration, Date, Text.

Availability depends on the report typeBaidu Analytics is report-type based. The report type is chosen when the data source is created and cannot be changed afterward. Each report type returns only its own metric set, so a metric listed below is available only in the report types that expose it. See Metric availability by report type for the exact mapping.

Users & sessions

Metric	Description	Data type	API Key
Visits	Number of visits (sessions) to the site.	Number	`visit_count`
Visitors	Distinct visitors (unique users).	Number	`visitor_count`
New visitors	Visitors recorded as first-time.	Number	`new_visitor_count`
New visitor rate	Share of visitors that were new.	Percentage	`new_visitor_ratio`
Unique IPs	Distinct IP addresses observed.	Number	`ip_count`

Traffic sources (Baidu promotion)

Metric	Description	Data type	API Key
Impressions	Times a promotion was shown.	Number	`show_count`
Clicks	Clicks on a promotion.	Number	`clk_count`
Cost	Promotion spend.	Currency	`cost_count`
Click-through rate	Clicks divided by impressions.	Percentage	`ctr`
CPM	Cost per thousand impressions.	Currency	`cpm`
Arrival visits	Visits that reached the site after a promotion click.	Number	`in_visit_count`
Arrival rate	Share of clicks that resulted in a site arrival.	Percentage	`arrival_ratio`

Engagement & behavior

Metric	Description	Data type	API Key
Page views	Total page views (PV).	Number	`pv_count`
Page view share	Share of page views attributed to the row.	Percentage	`pv_ratio`
Average pages per visit	Page views divided by visits.	Ratio	`avg_visit_pages`
Average visit duration	Mean length of a visit.	Duration	`avg_visit_time`
Average time on page	Mean time spent on the page.	Duration	`average_stay_time`
Bounce rate	Share of single-interaction visits.	Percentage	`bounce_ratio`
Entrances	Visits that started on this page (landing count).	Number	`visit1_count`
Outbound clicks	Clicks leading away from the page.	Number	`outward_count`
Exits	Visits that ended on this page.	Number	`exit_count`
Exit rate	Share of views on this page that were the last in the visit.	Percentage	`exit_ratio`

Conversions

Metric	Description	Data type	API Key
Conversions	Recorded conversion (transaction) count.	Number	`trans_count`
Conversion rate	Conversions divided by visits.	Percentage	`trans_ratio`
Average conversion cost	Promotion cost divided by conversions.	Currency	`avg_trans_cost`
Revenue	Revenue attributed to the report.	Currency	`income`
Profit	Profit attributed to the report.	Currency	`profit`
Return on investment	Return relative to promotion cost.	Ratio	`roi`

Metric availability by report type

The report type selected at creation determines which of the metrics above are returned. The eventAnalysis report type has no fixed metric set — its metrics are defined per request.

Report type	UI label	Available metrics
`websiteOverviewTrendData`	Website Overview (Trend Data)	`pv_count`, `visitor_count`, `ip_count`, `bounce_ratio`, `avg_visit_time`, `trans_count`
`searchTerm`	Search Term	`pv_count`, `pv_ratio`, `visit_count`, `visitor_count`, `new_visitor_count`, `new_visitor_ratio`, `ip_count`, `bounce_ratio`, `avg_visit_time`, `avg_visit_pages`, `trans_count`, `trans_ratio`
`visitedPages`	Visited Pages	`pv_count`, `visitor_count`, `ip_count`, `visit1_count`, `outward_count`, `exit_count`, `average_stay_time`, `exit_ratio`
`visitedDomainName`	Visited Domain Name	`pv_count`, `pv_ratio`, `visit_count`, `visitor_count`, `new_visitor_count`, `new_visitor_ratio`, `ip_count`, `bounce_ratio`, `average_stay_time`, `avg_visit_pages`
`searchEngines`	Search Engines	`pv_count`, `pv_ratio`, `visit_count`, `visitor_count`, `new_visitor_count`, `new_visitor_ratio`, `ip_count`, `bounce_ratio`, `avg_visit_time`, `avg_visit_pages`, `trans_count`, `trans_ratio`
`geographicalDistributionByCountry`	Geographical Distribution (by country)	`pv_count`, `pv_ratio`, `visit_count`, `visitor_count`, `new_visitor_count`, `new_visitor_ratio`, `ip_count`, `bounce_ratio`, `avg_visit_time`, `avg_visit_pages`, `trans_count`, `trans_ratio`
`geographicalDistributionByProvince`	Geographical Distribution (by province)	Same as Geographical Distribution (by country)
`websiteOverviewCommon`	Website Overview (Source, Search Terms, Entry Page, Visited Pages)	`pv_count`, `visit_count`, `visitor_count`, `new_visitor_count`, `ip_count`, `bounce_ratio`, `avg_visit_time`, `avg_visit_pages`
`trendAnalysis`	Trend Analysis	`pv_count`, `pv_ratio`, `visit_count`, `visitor_count`, `new_visitor_count`, `new_visitor_ratio`, `ip_count`, `bounce_ratio`, `avg_visit_time`, `avg_visit_pages`, `trans_count`, `trans_ratio`, `avg_trans_cost`, `income`
`promotionMethod`	Promotion Method	`show_count`, `clk_count`, `cost_count`, `ctr`, `cpm`, `pv_count`, `visit_count`, `visitor_count`, `new_visitor_count`, `new_visitor_ratio`, `in_visit_count`, `bounce_ratio`, `avg_visit_time`, `avg_visit_pages`, `arrival_ratio`
`baiduPromotionTrends`	Baidu Promotion Trends	`pv_count`, `pv_ratio`, `visit_count`, `visitor_count`, `new_visitor_count`, `new_visitor_ratio`, `ip_count`, `bounce_ratio`, `avg_visit_time`, `avg_visit_pages`, `trans_count`, `trans_ratio`, `avg_trans_cost`, `income`, `profit`, `roi`
`eventAnalysis`	Event Analysis (事件分析)	Dynamic — defined per request; no fixed metric set

Breakdowns

How to read the columnsCode style like _date is the literal field name Adriel stores for the breakdown. Italic text describes a value that Adriel derives rather than reading it directly from a single API field.

The primary row breakdown is set by the report typeBeyond the date breakdown, each report type produces rows grouped by its own entity (for example, one row per search term, per visited page, or per province). Because the report type is fixed at creation, that primary breakdown is fixed for the data source and cannot be swapped for another entity without recreating it.

Date

Breakdown	Description	API Key
Date	Daily date of the row, normalized to `YYYY-MM-DD`.	`_date`

Report entity

The row entity is determined by the report type selected at creation. The table keys each breakdown by its report type; the underlying entity-column name varies per report type.

Breakdown	Description	API Key
Search term	One row per search keyword that led to the site.	`searchTerm`
Visited page	One row per page on the site.	`visitedPages`
Visited domain	One row per referring or visited domain.	`visitedDomainName`
Search engine	One row per search engine source.	`searchEngines`
Country	One row per country of the visitor.	`geographicalDistributionByCountry`
Province	One row per province of the visitor.	`geographicalDistributionByProvince`
Promotion method	One row per Baidu promotion method.	`promotionMethod`
Website overview (trend)	Rows summarize the site over time (no entity split).	`websiteOverviewTrendData`
Trend analysis	Rows summarize the site over time (no entity split).	`trendAnalysis`
Website overview (common)	Rows depend on the selected collection (source, search terms, entry page, or visited pages).	`websiteOverviewCommon`
Event	Rows depend on the event breakdowns defined for the report.	`eventAnalysis`

Time & date grouping

Adriel groups Baidu Analytics’ daily results into these standard views for reporting alongside other connectors.

Breakdown	Description
Auto Time Breakdown	Automatically picks the best granularity for the date range.
Daily	Split reports by day.
Weekly (Start: Mon)	Split reports by week, starting Monday.
Weekly (Start: Sun)	Split reports by week, starting Sunday.
Week Date Breakdown	Split by the start date of each week.
Week Number Breakdown (W#)	Split by week number in the year.
Monthly	Split reports by calendar month.
Quarterly	Split reports by calendar quarter.
Yearly	Split reports by calendar year.

Limitations

Based on the current connector, Baidu Analytics has:

A fixed report type per data source — the report type is chosen at creation and cannot be changed later; a different report requires a new data source.
A single site per data source — selecting multiple sites during setup provisions one data source per site.
A 180-day history cap — only the last 180 days are backfilled on first connection; earlier data is unavailable.
A 30-day rolling refresh — only the most recent 30 days are re-fetched on each scheduled run.
Daily granularity — the source returns daily data; sub-daily rows are not provided.
Manual reconnection on token expiry — the connection does not auto-refresh; an invalid or revoked access token disconnects the data source and requires re-entering credentials.
Beta status — the connector is presented in the product as “Tongji (beta).”

API references

Baidu Tongji API (百度统计 API)

​Introduction

​Data refresh strategy

​Architecture data

​Reports data

​Architecture levels

​Date range limits

​Metrics

​Users & sessions

​Traffic sources (Baidu promotion)

​Engagement & behavior

​Conversions

​Metric availability by report type

​Breakdowns

​Date

​Report entity

​Time & date grouping

​Limitations

​API references

​See also

Introduction

Data refresh strategy

Architecture data

Reports data

Architecture levels

Date range limits

Metrics

Users & sessions

Traffic sources (Baidu promotion)

Engagement & behavior

Conversions

Metric availability by report type

Breakdowns

Date

Report entity

Time & date grouping

Limitations

API references

See also