Google Maps Scraper

Hi, could you please share the old and new run? I'll take a look.

If you don't want any unrelated bussineses, try to use category filter. You can also limit the number of places per search term.

Hi,

When I run the actor I am specifying the business types in the search terms and defining the area by JSON code.

When reviewing the output of the most recent run (see attached) over 75% of the business categories scraped (Variable: ‘categoryName’) where not even closely related to the business categories included in my search terms (see attached).

This resulted in over 50% of the businesses scraped coming from these categories that are not related to my search terms.

Are you able to see why this is happening and if there is any way I can avoid this in the future?

Hi Ondrej,

Do you have any update on this?

Thanks

Hi, thanks for the data and I apologize for late response.

The reason why you're getting more results is probably because we released a new version of the actor with improved search capabilities - it uses "search this area" feature (see screenshot) that allows us to find more places.

Regarding why you see many unrelated categories:

1. categoryName contains only the main category of a place. All place's categories are stored in categories. Take for example this place(629th row in dataset): it's main category is Gas station, but the other categories are ATM, Cafe and restaurant. So these places should be fine.
2. sometimes Google Maps gives us results that it thinks is "close enough" to what we want, so it gives us places like this (it's categories are Motel and Hotel).
3. you have enabled scraping directories ("scrapeDirectories": true) - this tells the actor to scrape places that are inside another places (e.g. stores, restaurants, etc. in malls). But in your run, actor found this restaurant that has ATM inside, so actor scraped that as well.

You can get rid of the places from 2. and 3. point by adding category filters. Based on your list of search terms, I'd use the following categories:

1"categoryFilterWords": [
2        "restaurant",
3        "bar",
4        "coffee shop",
5        "cafe",
6        "ice cream shop",
7        "sandwich shop",
8        "pizza delivery service",
9        "pizzatakeaway",
10        "juice shop",
11        "brewery",
12        "diner",
13        "steak house",
14        "dessert shop",
15        "bubble tea store",
16        "tea house",
17        "pub",
18        "food court",
19        "bistro",
20        "brewpub",
21        "acai shop",
22        "chinese takeaway",
23        "sushi takeaway",
24        "art cafe",
25        "creperie"
26    ],

Hi Ondrej,

Thanks for getting back to me with a detailed response. I can see Zuzka Pelechová has closed the issue but I still have further questions.

With regards to the proposed solution you have provided of applying category filters – does each one have to be selected manually or can I paste them in? If I apply those filters will I still capture all of the categories in my search terms? That I essential to me.

Thanks for your explanation, understanding the scrape better now I have reviewed the extract again. I have attached the workings file – just extracted the categories, CID and URL for reference.

When looking at all categories associated with a business I found;

• Of the 3,627 Businesses scraped, only 1,458 of them (40%) have a category that matches the search terms
• Of the 3,627 Businesses scraped, only 1,859 of them (51%) have a category that matches the search terms or is relevant and close to the search terms
• Of the 3,627 Businesses scraped, only 3,093 of them (84%) either match, are relevant, or are located inside somewhere whereby ‘scraping directories ("scrapeDirectories": true)’ captures them.

So there are 579 (16%) businesses with 460 different business categories that are either not relevant or located in a space that could be scraped as part of the ‘("scrapeDirectories": true)’ function. Is there any other way to not capture these (and not spend my data allowance on them)?

If I turn off the ‘scraping directories ("scrapeDirectories": true)’ is it done here and will this remove all the listings from unrelated categories? [cid:image002.png@01DA97DE.6A23DCC0]

If not, I either need to build a work around on my side or change scraping tools

Thanks for your help

Hi there,

I’m just wondering if someone is still looking at this?