Choose a source
Upload an IP file or select a period of external advertising reports.
How it works
These are alternative input paths. CSV/ZIP/GZIP uploads feed file processing; Celery download/ETL chains supply the report branch.
Advertising traffic segmentation and export service
A traffic-analysis pipeline that combines advertising reports, IP context and completeness-aware filtering.
Project overview
What the service does
A traffic-analysis application for advertising operations. Users either upload a file of IP addresses or select a date range from imported advertising reports, then filter by geography, IP range, click or impression counts, or deposit counts. The result is a downloadable segment; one workflow can also upload a list to the configured advertising service.
Large advertising reports are hard to inspect manually. Operations need reproducible criteria for identifying IP addresses and ranges, and a way to turn those decisions into usable exports.
Who uses it
Traffic analysts and advertising operations teams who review large reports and prepare repeatable IP or audience segments.
Input
- An uploaded CSV/ZIP/GZIP file or a selected reporting period.
- Geography, IP/CIDR, click/impression and deposit filters.
- Imported daily advertising reports for report-based queries.
Output
- A CSV file containing the selected IP segment.
- Task progress and errors visible in the interface.
- For a configured workflow, a list upload to an external advertising service.
System requirements
- Daily external reports must be loaded into analytical storage.
- Filters depend on data completeness for the requested interval.
- IP and geographic context support traffic interpretation.
- Results can be downloaded or optionally sent to an external blacklist.
01 / How the service works
IP file or report → traffic filters → selected segment
Process mapSolid line: normal flow · dashed line: retry or return
- INPUTUpload IP file
CSV, ZIP or GZIP input
ContinueNormalize addresses - STEPNormalize addresses
Prepare IP values and ranges
ContinueApply file filters - SERVICEApply file filters
GeoIP and set/range operations
ContinueDownload segment - RESULTDownload segment
CSV result for the operator
- INPUTChoose report dates
Imported advertising data
ContinueLoad and prepare reports - SERVICELoad and prepare reports
Daily import into analytical storage
ContinueAll dates present? - DECISIONAll dates present?
Missing days stop an incomplete lookup
completeRun traffic filtersmissing datesShow missing dates - SERVICERun traffic filters
IP, user-agent, click/impression counts or deposits
ContinueDownload or send list - RESULTDownload or send list
CSV, with optional configured upload
- RESULTShow missing dates
Ask for a complete reporting interval
02 / System architecture
Main components and their roles
Prepare the input
Each source is prepared for its own filtering method.
How it works
The file path normalizes IPv4 and mapped IPv6 addresses. The report path loads rows into ClickHouse staging tables and updates the relevant date partitions.
Check readiness
Task state tracks the input; report-based filters also check date coverage.
How it works
PostgreSQL stores task/file metadata. Missing-date detection applies to analytical report lookups, not as a mandatory step for every file upload.
Apply segment rules
The selected workflow produces a segment using geography, event counts or address ranges.
How it works
For an uploaded file, GeoIP2 adds location data and the service applies IP/range filters. For advertising reports, ClickHouse filters by IP, user-agent, clicks, impressions, deposits or CIDR.
Inspect progress
The interface shows execution progress, output state and errors.
How it works
PostgreSQL stores each task state and file metadata; the interface uses them to show progress and the result of a filter.
Export the result
The result becomes a downloadable list or an optional platform blacklist update.
How it works
The service creates a CSV file; a configured workflow can also send the selected list to the advertising platform.
03 / Implementation details
Engineering decisions
Separate operational and analytical data
PostgreSQL stores application state while ClickHouse serves report-scale analytical access.
Gate incomplete intervals
Data availability becomes a condition for running a meaningful filter.
Close the operational loop
A filter produces an export that operators can apply in their advertising workflow.
Software development for business
Have a project in mind?
Tell me what the application should do, how the team works today and which requirements or deadlines matter.