DaData.ru calculates the coordinates, and the address parser wags its tail
DaData.ru - a service of automatic verification and correction of contact data (name, address, phone number, email).
In the courtyard in November, and at DaData.ru - new features:
Once the operator of the Tula call center accepted the order and drove the address “Moscow Nagornoye 1” instead of “Moscow Nagornaya Street 1”. Then the address was automatically driven through Yandex and received the geo-coordinates of the Nagorny highway in Khimki. The parcel left for Khimki, and they took it out with great difficulty.
To reduce the number of such stories, we added geo-coordinates to the API . Pass the address in one line, and we will analyze by KLADR and put down the geo-coordinates.
We search carefully - we return the street coordinates if a house is not found. And the coordinates of the city center, if the street is not found.
We always indicate the accuracy with which the geo-coordinates of the object were found - is it a house, street or settlement (see quality codes ).
We use OSM .
Coverage of houses:
Street Coverage:
Gradually we plan to supplement the data on geocodes to increase coverage.
The online store making SMS deliveries, the API will determine the time zone at the delivery address.
For those whose CMS does not allow you to enter the address in one line, they made granular hints for the addresses .
It used to be inconvenient to enter a company or address on a smartphone.
And now even the owner of a 3.5-inch phone will fill in the details of the company in 10 seconds (we checked).
TIN in a separate line and clearly visible. And the address does not clog the space, only the region is left of it.
Last year, we actively reworked the algorithm for recognizing mailing addresses. A year of heated debate between developers and testers on the topic "how to parse addresses correctly." A year of constant code fixes and daily monitoring of automated tests.
And here it is - a new address filter. Call it through the API or process the files. Tested for 50 million autotests.
Many thanks to those who wrote errors on the forum!
Special thanks to the authors of the topics:
See statistics on processed records in your account.
Nikolay Sotsky from InstaDev made an excellent module for calling prompts from iOS: https://github.com/InstaDev/IDDaDataSuggestions .
We will be glad to questions on the forum !
In the courtyard in November, and at DaData.ru - new features:
- definition of geo-coordinates,
- SMS in the daytime,
- new address algorithm
- optimization for smartphones,
- statistics in your account
- module for iOS.
Do not let the parcel go to Khimki instead of the center of Moscow
Once the operator of the Tula call center accepted the order and drove the address “Moscow Nagornoye 1” instead of “Moscow Nagornaya Street 1”. Then the address was automatically driven through Yandex and received the geo-coordinates of the Nagorny highway in Khimki. The parcel left for Khimki, and they took it out with great difficulty.
To reduce the number of such stories, we added geo-coordinates to the API . Pass the address in one line, and we will analyze by KLADR and put down the geo-coordinates.
We search carefully - we return the street coordinates if a house is not found. And the coordinates of the city center, if the street is not found.
We always indicate the accuracy with which the geo-coordinates of the object were found - is it a house, street or settlement (see quality codes ).
We use OSM .
Coverage of houses:
- Moscow - 45%
- other million-plus cities - 35%,
- the rest of Russia is 10%.
Street Coverage:
- Moscow - 90%
- other million-plus cities - 60%,
- the rest of Russia is 25%.
Gradually we plan to supplement the data on geocodes to increase coverage.
Send SMS on time
The online store making SMS deliveries, the API will determine the time zone at the delivery address.
Tell me by region, city, street and home
For those whose CMS does not allow you to enter the address in one line, they made granular hints for the addresses .
Tell me on smartphones
It used to be inconvenient to enter a company or address on a smartphone.
And now even the owner of a 3.5-inch phone will fill in the details of the company in 10 seconds (we checked).
TIN in a separate line and clearly visible. And the address does not clog the space, only the region is left of it.
Process addresses through a new smart filter
Last year, we actively reworked the algorithm for recognizing mailing addresses. A year of heated debate between developers and testers on the topic "how to parse addresses correctly." A year of constant code fixes and daily monitoring of automated tests.
And here it is - a new address filter. Call it through the API or process the files. Tested for 50 million autotests.
Many thanks to those who wrote errors on the forum!
Special thanks to the authors of the topics:
- Petrovsko-Razumovsky passage MSC
- Error in determining the area from the address for the Republic of Bashkortostan
- Prioritize regional centers
- 6 Novopodmoskovniy lane transfers to Novopodmoskovniy 1st
- How should the address Moscow, 47 km of Moscow Ring Road, be understood by KLADR?
- Doesn't make out Pilot Gritsevets street
- Incorrectly recognizes addresses in Rostov
- Error defining Richard Sorge Street
- Errors in determining Moscow addresses
Count money
See statistics on processed records in your account.
Turn on hints for iOS
Nikolay Sotsky from InstaDev made an excellent module for calling prompts from iOS: https://github.com/InstaDev/IDDaDataSuggestions .
We will be glad to questions on the forum !