public-roadmap
public-roadmap copied to clipboard
serpapi
[Google Trends API] Searches Returning Incorrect Results
A customer reported Google Trends searches are sometimes returning inaccurate results. I haven't been able to replicate it in the playground, but they shared several "Interest over Time" searches that contain 0 values. I don't see a specific pattern in the dates used.
Search Inspector: 1 | 2 | 3 | 4 Intercom
I had thought this might have been on Google's side and could be already resolved. However I was able to recreate it in the Playground, so it is still a problem:
Is this getting any attention? Seems like it just keeps getting reported without any resolution.
Moved this to urgent because of the high number of customers who have inquired about it.
However it's possible there won't be a fix for this. Pytrends experienced the same issue and concluded it was due to factors beyond their control:
https://github.com/GeneralMills/pytrends/issues/581
User shared a file with details of over 3,000 failed searches in case it is helpful for investigating:
is there any update at all on this?
Rendering the trends feature useless at this point.
Appreciate the hard work!
This is now completely inoperable. Its been two months since I flagged and there has been no meaningful update.
This is now completely inoperable. Its been two months since I flagged and there has been no meaningful update.
Sorry about that. There are unfortunately several technical issues both our side and Google side at the moment with Google Trends data. We are working on it though. And hopefully we will have fixes soon.
Data from Google Trends and SerpApi match and contain only positive values.
| Google Trends | SerpApi |
|---|---|
| https://trends.google.com/trends/explore?date=2023-03-04T15%202023-03-05T15&geo=US&q=nissan&tz=420 | https://serpapi.com/playground?engine=google_trends&q=nissan&geo=US&tz=420&date=2023-03-04T15+2023-03-05T15 |
Another customer reported the issue.
Another customer reported the issue.
Unfortunately, the Interest Over Time values in the response differ when we use a proxy, even with the exactly same URL and cookies as in the real browser. Without a proxy, values in the response are the same.
Additionally, the widget URLs we use in the code are also different.
URL we used:
https://trends.google.com/trends/api/widgetdata/multiline?tz=300&req=%7B%22time%22%3A%222023-12-01T11%5C%5C%3A26%5C%5C%3A59+2023-12-08T11%5C%5C%3A26%5C%5C%3A59%22%2C%22resolution%22%3A%22HOUR%22%2C%22locale%22%3A%22en-US%22%2C%22comparisonItem%22%3A%5B%7B%22geo%22%3A%7B%7D%2C%22complexKeywordsRestriction%22%3A%7B%22keyword%22%3A%5B%7B%22type%22%3A%22BROAD%22%2C%22value%22%3A%22sakura%22%7D%5D%7D%7D%5D%2C%22requestOptions%22%3A%7B%22property%22%3A%22%22%2C%22backend%22%3A%22CM%22%2C%22category%22%3A0%7D%2C%22userConfig%22%3A%7B%22userType%22%3A%22USER_TYPE_SCRAPER%22%7D%7D&token=APP6_UEAAAAAZXRPpFCno2qIeTJuya_mkdccGxdScUHH
Google Trends UI:
https://trends.google.com/trends/api/widgetdata/multiline?hl=en-US&tz=300&req=%7B%22time%22:%222023-12-01T11%5C%5C:26%5C%5C:59+2023-12-08T11%5C%5C:26%5C%5C:59%22,%22resolution%22:%22HOUR%22,%22locale%22:%22en-US%22,%22comparisonItem%22:%5B%7B%22geo%22:%7B%7D,%22complexKeywordsRestriction%22:%7B%22keyword%22:%5B%7B%22type%22:%22BROAD%22,%22value%22:%22sakura%22%7D%5D%7D%7D%5D,%22requestOptions%22:%7B%22property%22:%22%22,%22backend%22:%22CM%22,%22category%22:0%7D,%22userConfig%22:%7B%22userType%22:%22USER_TYPE_LEGIT_USER%22%7D%7D&token=APP6_UEAAAAAZXRPA7alxgSor3RXKAUT0IGPTaWyBiB6&tz=300
Hi @nicktba. Unfortunately, we don't have a proper solution to solve the issue yet. Our Engineers are trying to find the solution, but the issue seems more complex than anticipated.
Meanwhile, another customer reported the same issue.
Hi @edmarsj, unfortunately, we don't have an update on this issue at this time. Any fixes, workarounds, or decisions regarding the issue will be posted here. If you've contacted us via email or chat, we will also reach out to you there.
Thanks for your continued patience!
Another customer reported that when searching for "Compared Breakdown By Region" (data_type=GEO_MAP), the results are different than in an actual search in Google Trends.