Flickr Downloadr: Scraping Flickr Imagery, People Photostreams, and Group Collections
By Shalin Hai-Jew, Kansas State University
Figure 1: “Digital”-seeded Image Search on Flickr (within the “Flickr Downloadr” App)
In the Web 2.0 era, people have been sharing copious amounts of digital data via their social media accounts. Of the various types of shares, digital images have been some of the most portable and adaptable to other-than-original contexts. For example, capturing imagery related to particular historical and contemporaneous events, cultural practices, social phenomena, styles, public figures, and other topics may enhance teaching and learning, as well as research (such as content analysis based on user-generated contents).
The early challenge in such endeavors involves how to capture and scrape a large number of images to bulk-download images in useful digital formats. While there are some browser-based tools that enable data scraping from various dedicated sites and across sites, there are Web-browser-based tools that enable data extractions from specific social media platforms. These latter tool types are coded to the application programming interfaces (APIs) of the respective social platforms and so often enable a smoother user experience and sometimes a larger haul of imagery than might be possible otherwise. [Those who can deploy crawlers may be able to capture even larger amounts of imagery, but this requires additional skills.] It is also important to capture images in a viewable and (computer- and human-) accessable way. There are dedicated tools that enable access to one’s own account. On log-in, a person is able to access all of his / her images in a mass bulk download. In this case, what is desired is the ability to capture others’ imagery based on text-based searches (in a number of UTF-8 enabled languages).
For those who use Flickr, with 10 billion images shared by its 122 million users in 63 countries, there is a a fairly recent Google Chrome browser add-on that enables easy download of hundreds of target images at a time (apparently based on textual image file names, folk tags, and descriptions). This article describes how to download and use the free Flickr Downloadr from the official Google Chrome Store Extensions site.
Accessing and Installing Flickr Downloadr
To get started, it helps to have Google Chrome installed (for whatever platform the user is using). This app is compatible with web-enabled devices.
Figure 2: Adding Extensions to Google Chrome Web Browser
- Start the Google Chrome web browser.
- Click the “Customize and control Google Chrome” button indicated by a set of three vertical ellipses (three vertical dots) at the top right of the menu bar of the Google browser window.
- Click “More tools,” and in the dropdown menu, select “Extensions.”
- A list of installed extensions will appear in alphabetical order. Scroll to the bottom, and click the link “Get more extensions.” The link will connect you to the Chrome Web Store.
- In the “Search the store” field, type “Flickr Downloadr.” Click return.
- There are several similar extensions (DownFlickr-Flickr Downloader,” “Flickr downloader,” and “Flickr Photos Download.” On the Web, there are actually many others. The one mentioned here is Flickr Downloadr by brain.no. (The developer is Sondre Bjellås.)
- Once the add-on is added, it may be directly launched with the green “Launch app” button. Or it may be started in the Search box at the bottom left of the screen.
- The Flickr Downloadr page looks like the following:
Figure 3: Flickr Downloadr App Page
With this pop-up window, people may consider an overview of the add-on. They can read verified user Reviews (based on verified identities from Google + accounts). They may read Support messages by other users. [As a small aside, this add-on enables access to the Apollo Image Gallery as part of the “Project Apollo Archive,” hosted on Flickr. There is no direct relationship between the maker of the tool and Yahoo’s Flickr.
Three Types of Imagery
The Flickr Downloadr tool enables access to three types of data from Flickr:
(1) imagery (collections seeded from text searches based on Boolean search features, like “and,” “or,” “not,” “(),” and so on, and including keywords, hashtags, and other texts),(2) people (their photostreams), and(3) groups (their photostreams).
(1) includes ad hoc image scraped based on the seeding terms. (2) includes photo collections based on "personality" or the individual behind the images. (3) includes photo collections created by a group of individuals, often with shared photography-based (or other) interests.
(1) Capturing Flickr Imagery
The steps to acquiring imagery generally goes as follows:
- Start the Flickr Downloadr desktop application.
- In the field at the middle, type in the text search for the desired photo, people, or group.
- Click return.
- At the top left, choose “Photos,” “People,” or “Groups.”
- Click “Load more…” to capture more imagery.
- If this particular session is not desirable, just click on the three lines at the left to access the menu, and the text search box appears again at the top left.
- At the bottom right, three are two menus. One menu enables the selection of proper licensure: “Any license,” “Creative Commons,” and “CC (Commercial use)”.
- To the right of that, the menu enables filtering of imagery by any one (at a time) of the following: “Date uploaded,” “Date taken,” “Interesting,” and “Relevant.”
- It helps to experiment with the various types of filters to see how the image selections change. The “Any license” seems to enable the broadest range of image types in terms of licensure restrictions. The “Interesting” does seem to pull out the images with the most unusual shapes and contexts. The “Date taken” and “Date uploaded” filters seem to result in a wide range of images, not all directly related to the topic.
- Click the 4-cell b/w button at the bottom of the screen. This automatically selects all the images found by the Flickr Downloadr. Once the desired images are selected, click the ones which are not desired to deselect.
- At the bottom, there is an arrow pointing right. Click that button to start the download.
- The next screen requires the user to select the file or folder that the images will download to.
- Click the down arrow at the bottom right to start the download process.
- Check the files to make sure that they are correctly downloaded. (The file names are the ID names of the imagery on Flickr.)
Figure 4: “Any Licensed” and “Interesting” Chocolates on Flickr
By default, the Flickr Downloadr settings are automatically set on SafeSearch. This can be mitigated somewhat by putting it on SafeSearch moderate or lifted altogether with SafeSearch off. To access the Settings, users click on the Settings at the bottom left.
Figure 5: Flickr Downloadr Settings Screen
At the bottom left, the Log-in is to Flickr.
(2) Sampling People’s Photostreams
To access Flickr users’ photostreams, users can click on the “People” selection at the top left of the app window. What follows is a hand-off to the Flickr site. If the search term is sufficiently disambiguated to locate to a person, the app will pop up “Found user.”
In this case, the account for the maker of Flickr Downloadr, Mr. Sondre Bjellås, was selected. This account is located at https://www.flickr.com/photos/sondre. This account (SondreB) contains almost 2,000 photographs; it has 74 followers and 81 following. The user joined back in October 2005. Some of the images relate to Flickr Downloadr, people, landscapes, and other topics.
Figure 6: People -> Profile -> Photostream (on Flickr)
Once a user account has been identified, one may access the following aspects of their (public) account: Photostream, Albums, Favorites, Galleries, and their actual Profile.
(3) Accessing Groups’ Imagery Holdings
A third information type that is accessible with the Flickr Downloadr is group imagery holdings. A recent article about Flickr statistics puts the number of groups at about two million [“flickr Stats (November 2016)”], but a fair number of these have lapsed into silence. When pulling up groups, there is a helpful feature in Flickr to leave off inactive groups, which simplifies searches for active groups.
In this app, what is initially pulled up are some thumbnails of group image identifiers…and then as one hovers over each of the icons, one can see more information about that group. At the bottom right of each image is a listing of the number of related images to that group account. (A group is a cluster of Flickr users with shared interests around a topic, and these groups include declared members. Groups may be public or private, but in terms of accessing group data, only public groups may be accessed.)
For this extraction, the Libraries and Librarians account was used. This group came up in a listing of those groups related to “technologies…” This Libraries and Librarians group contains 4,657 members and includes some 53,000 photos. This group also has 102 discussions around a range of topics. The Flickr account handle is “libraryman.”
Figure 7: “Libraryman” Search Results in Groups (55,009 Images)
In terms of image accesses, because there are varying levels of licensure with images in groups, extracting group imagery often results in successful image downloads but also “error messaging” that reads as follows: “Downloaded 58 photos successfully and skipped 235 photos due to license or permission restriction.”
Finally, there are all sorts of creative approaches to using words and symbols and numbers to elicit imagery. The examples in this short work are fairly limited, and it's good to go well beyond the limited ideas here.
Figure 8: Search Variations in Flickr (using Flickr Downloadr)
Note: Sometimes, using applications can result in limits to the accessed data. For example, using Flickr Downloadr to access "group" data on Flickr resulted in one set of groups while going directly to Flickr resulted in a different and much larger set of groups. Further, the second data extraction directly from Flickr resulted in a much larger list of active groups on the image- and video-sharing platform. This is to say that those who are using applications for research should be aware of various limitations...and should approach data extractions in multiple ways to find out which approach(es) is(are) optimal.
About the Author
Shalin Hai-Jew works as an instructional designer at Kansas State University. She may be reached at firstname.lastname@example.org.
|Previous page on path||Cover, page 20 of 27||Next page on path|