Dehumanising the homeless through language

Post 6 by Alice Stollery

Web Scraping Tools

Twitter has become both a social and professional platform allowing dissemination of anything from random thoughts, ignorance and pointless memes to breaking news and public opinion. With 313 million active global users recording and sharing their thoughts, feelings and experiences, twitter is a real-time source of information and has become a medium in which people can keep up with those they know and those they don’t. The 140-character limitation on posts makes it a perfect tool for researching public opinion and scraping the platform for data, without being overwhelmed.

Web Scraping Process

Twitter archiver was my main source when scraping the web, as it allows you to create a search rule and, unlike other tools such as twitter advanced search or facebook, it collates the results into a single excel spreadsheet. This feature became invaluable in gaining insights from the results as it allowed me to sift through the information using specific terms, further refining my results. Insights often came as a result of tangents, originating in previous search results.

FullSizeRender 40
A simple visualisation of my process of navigating web scraping and twitter archiver. I found I went off on tangents from previous results.

Not being a twitter user myself, I stumbled at the start. I created a twitter account and downloaded twitter archiver and linked it to my gmail account. Initially, I ran searches that were very specific, in an attempt to find information on stereotyping and the technological divide within homelessness. I set the location to Sydney, however, this failed as I did not receive a single result. So, I changed the location and widened the search to Australia. Again, I was being too specific with the number of words I was using and twitter archiver came back with nothing.

Screen Shot 2016-09-05 at 23.26.30
A simple visualisation of my process of navigating web scraping and twitter archiver. I found I went off on tangents from previous results.


I then decided to change my approach and create a simple search of ‘homeless’ and ‘stereotyping’ which produced 14, 931 results. I then used the search tool to search for key words from the group issue mapping class within the data set. This helped me to break down the vast amount of information and to see what people were saying about issues within homelessness. This became a very interesting process. The initial words I searched within the results were quite basic as I was still getting use to using the software. However, they still provided insights and I have listed the most frequent words below.

Screen Shot 2016-09-06 at 00.08.42
A simple visualisation of my process of navigating web scraping and twitter archiver. I found I went off on tangents from previous results.

Interestingly, homeless men were referred to 4,279 times compared to only 251 times for homeless women. This is in line with the statistics revealed through one of my earlier posts that stated 82% of Sydney’s homeless are male and only 17% are female. As these posts were not Sydney specific, it is interesting to see that these statistics may also be a good indication of the situation in other parts of the world. I am curious as to why the numbers of homelessness between sexes differs so significantly. And can begin to understand why previous sources outlined a lack of services tailored to homeless women. I can only assume that this is due to less demand for them..? I could look further into the specific causes of male homelessness, contrast it to female homelessness and see if there is an opportunity to intervene with my design response.

Another observation was that the word ‘help’ only appeared in 1,137 posts out of 15,000. Looking further into these posts I found, ‘help’ took different forms. Some were genuine posts about helping the homeless with information of individuals lending a hand or community projects. It was really interesting to see how others were approaching and tackling the issue on a personal level.

genuine help
A simple visualisation of my process of navigating web scraping and twitter archiver. I found I went off on tangents from previous results.

Other twitter users made genuine offers help. In his case twitter was used as a form of communication to reach people in a particular area to help assist in helping those in need. This highlighted the possibility of technological responses that act as a connection between those with something to offer and those in need. Service design could bridge the gap between say, small businesses with food or accomodation available and the homeless.

Screen Shot 2016-09-06 at 00.28.26Screen Shot 2016-09-06 at 00.28.35

While each of the above insights were interesting, perhaps the most interesting insight was not only the presence of stigma in these posts but stigma as a result of the frequent misuse of the word homeless. While there were 15,000 posts that included the word homeless, I found that more often than not people were using it as part of casual conversation, to describe their lack of dress sense or effort invested in the appearance of friends or celebrities.

Screen Shot 2016-09-05 at 23.46.30
A simple visualisation of my process of navigating web scraping and twitter archiver. I found I went off on tangents from previous results.

Screen Shot 2016-09-06 at 20.57.23

It seems that, no matter how hard I try to broaden my understanding of the issue, I always seem to arrive back at stigma, perceptions of the homeless and ignorance towards their circumstances.After seeing these tweets, I decided to look into how the misuse of language is hindering our ability to tackle homelessness. As a result, I then ran a search on the term hobo to understand how often it was being used within the twitter sphere. This search found 10,732 results with the term hobo included in it.

“RT @emmaabel_: I am going to try and make myself look decent tmmr and not like a hobo”

“@Lizbeth_923: First Date And I Look Like A Hobo 🙂”

I believe the ubiquitous misuse of language surrounding homelessness is dehumanising the homeless, and ultimately, taking away from the issue. This is not only present in social conversations and personal online interactions, but is also reinforced by the fashion industry as seen below with the release of the ‘hobo’ bag.

Screen Shot 2016-09-06 at 00.59.19Shockingly, the term hobo is also being used as a design response to the issue, evident in the iHobo app. The virtual pet app that puts a homeless person in your pocket for you to feed and take care of. Forget to feed your hobo and he dies or runs off to get drugs. I think using gaming to tackle peoples perceptions is an interesting idea but I can’t help but feel this approach is distasteful, to say the least, and is reinforcing the publics stereotypical and often negative perceptions on the issue, not to mention dehumanising those in need.

Screen Shot 2016-09-06 at 18.50.39Screen Shot 2016-09-06 at 18.51.11Screen Shot 2016-09-06 at 18.52.20

To further investigate this area, I could also try searching the term ‘tramp’ and other common names used to describe the homeless. The above tweets highlight a severe lack of empathy among the general population for those suffering from homelessness. Homelessness does not seem to be a topic that people are talking passionately about. The homeless have fallen by the wayside and we are all so desensitised to the issue that homelessness has become commonplace in daily language for all the wrong reasons.

Design Response

Design responses could enable change in this area, and I would like to focus on the role language plays in the issue. I think there are already a number of individuals and NGO’s working to directly help the homeless so I would like to instead create a design response that tackles the wider issue and aims to influence the views people have of the homeless. Tackling this wider problem of perception, assumptions and language would aim to influence the common vernacular  rather than direct action on a smaller scale. This would hopefully result in a knock on effect, creating empathy and engagement among the wider population, to ultimately generate positive outcomes on a wider scale.

In terms of it’s form, I could create a twitter bot that calls people out on their misuse of particular words. However I think this would breed hostility rather than empathy. Language will be an important element and in order to generate a feeling of empathy I think the design would be suited to a poetic response that encompasses feelings and a the contradiction of meanings.

In an attempt to reveal relationships between language and the the number of homeless people, I could visualise the frequency of misuse of vital words. Perhaps I could plot the locations of these misuses and correlate this data with the number of homeless people in that particular area to see the relationships between the two and to discover how local attitudes affect the issue. However I am hesitant to do that as I do not think data will generate an empathetic response in the way that I am hoping.

Five Point Summary

  1. Twitter and twitter archiver are both very effective tools in scraping the web for data to understand how the wider population are feeling towards homelessness.
  2. It is important to remain open to outcomes outside your initial understanding. I went into this process with a focus on stigma and the technological divide, yet ended up delving further into the role language plays in creating barriers to a solution.
  3. Those offering food, services or accommodation on a personal level have great difficulty in finding the right people to help. Perhaps a service could be designed to bridge this disconnect.
  4. More often that not, conversation around homelessness is not referring to the issue at all and is used more so in casual conversation to describe appearances. The misuse of homeless terminology is rife among the online community and has seeped into the common vernacular ultimately resulting in a lack of empathy towards sufferers.
  5. Homelessness does not seem to be a topic that people are talking passionately about. The homeless have fallen by the wayside and we are all so desensitised to the issue that homelessness has become commonplace in daily language for all the wrong reasons.


AlaynaaaMarie. 2016, ‘I just added this to my closet on Poshmark: Michael Kors Julian Chain Exotic Hobo’, Twitter post, 2 September, viewed 2 September 2016,>.

Alex. 2016, ‘I look like a hobo walking to class but idgaf’, Twitter post, 1 September, viewed 2 September 2016,<>.

Bryant, M. 2010, iHobo app puts a homeless man in your pocket, The Next Web, viewed 31 August 2016,<;.

Dan Kenny, 2016. ‘Can’t think of any but I’ll happily help give it out in the streets after service, be done 11:30ish’, Twitter post, 2 September, 2 September 2016,<>.

Dream Closet. 2016, ‘Always torn between getting ready and looking cute or being lazy and looking like a homeless person’, Twitter post, 2 September, viewed 2 September 2016,<>.

Emmabel. 2016, ‘I am going to try and make myself look decent tmmr and not like a hobo’, Twitter post, 1 September, viewed 2 September 2016,<>.

EWapo. 2014, ‘I’m playing a game called iHobo where you look after a tramp and I’m legit checking up on him every 5 minutes, I’m here for you trampy’, Twitter post, 10 January, viewed 3 September 2016,<>.

FryGuy84. 2015, ‘who remembers that app iHobo when you had to look after a tramp like a tamagotchi’, Twitter post, 29 January, viewed 3 September 2016,<>.

Hungry Hoss, 2016. ‘At a wedding in Brighton & there’s a shit load of food left over… does anyone know a homeless shelter or similar who’d make use?’, Twitter post, 2 September, viewed 2 September 2016,<>.

Lizbeth_936. 2016, ‘First date and I look like a hobo’, Twitter post, 1 September, viewed 2 September 2016,<>.

NinjaRonnie. 2013, ‘Shit I forgot about my hobo’, Twitter post, 9 June, viewed 2 September 2016,<>.

Pissedoff99. 2016, ‘You ignorant description kind of shows how insensitive people involved with fashion industry can be with social issues’, Twitter post, 3 September, viewed 3 September 2016,<>.

Prank Videos, 2016. ‘He gave his ‘winning’ lottery ticket to a homeless man, gave the store owner the money to pretend its real, amazing’, Twitter post, 11 March, viewed 3 September 2016,<>.

Smith, C. 2016, Twitter Statistics and Facts, DMR Stats and Gadgets, viewed 2 September 2016,<>.


Twitter Archiver for collecting data

Blog post 6. Scraping the web for data

Written by Hyunjoung You


As Media Access Australia (n.d) states:

Twitter is a popular social networking tool that allows users to send a short, mostly text-based message up to 140 characters long known as a ‘tweet’. These tweets are then published online and can be publicly viewed. Twitter users can post their own tweets, follow the tweets of other users or contribute to a wider online discussion based on a particular topic or event.

Twitter is fast personal communication. People can share personal insights on something with other people. Moreover, they can follow the celebrities and send feedback on any events such as a live television show. It is also commonly referred to as a short web log (blog). Social Media News Australia reported that Twitter becomes Australia’s most popular social media microblogging tool with approximately 2.8 million unique visitors in Australia and over 300 million users worldwide in the early of 2016.

My research process 

Screen Shot 2016-09-21 at 12.44.54 pm.pngMy specific topic is the association between sedentary lifestyle and obesity. Therefore, I searched using keyword ‘Obesity’, ‘Fat’, ‘Sedentary’, and ‘Lifestyle’ at first.

Screen Shot 2016-09-21 at 12.54.14 pm.png
Twitter Archiver Research 1 – #obesity #fat #sedentary #lifestyle

The data that came out on the list was exactly same as what I though about. However, as you can see the above screen shot, only one tweet showed since I used too specific keywords. I realized that I needed to use more general and suitable words to collect useful data.

Screen Shot 2016-09-21 at 1.14.30 pm.png
Twitter Archiver Research 2 – #obesity #fat

This is the result by researching using keywords of ‘obesity’ and ‘fat’. I could receive lots of personal insights about obesity, but it was hard to find the information what I looked for because keyword was so broad to bring about specific data. Nevertheless, there were few results were related to my topic. After this, I searched using keywords ‘lifestyle’ and ‘office’ as well; however, it was not enough to gather useful data. Hence, I moved on Twitter search engine.

Screen Shot 2016-09-21 at 4.50.38 pm.png

twitter search 1.pngtwitter search 2.png

As you can see the above images, I typed three words ‘fat’, ‘sitting’, and ‘office’, which are more related to my topic. Many tweets came out, and they all indicated that sedentary work made them being fat. It shows that many people already recognize sedentary lifestyle is associated with obesity, but all tweets were their feelings about being fat like sad or anger. There were no any solutions or ideas for that issue.


While I scraped data via Twitter Archiver and Twitter search engine, I found how they were useful tool to discover information what I looked at using simple keywords. Twitter Archiver offered the list, which included the keywords I typed. It helped me to recognize what kinds of issues people share and discuss nowadays. Also, it provided wider knowledge that is related to obesity issue. Overall, I could have a look different personal insights and opinions about specific issues. It is really good to know them as a designer because we have responsibility to act for people needs and build the solutions to solve problems. Therefore, it is appropriate tool to scrape data to understand specific issue and personal insights.



Media Access Australia, n.d. ‘Twitter’, viewed 4 September 2016, <;

Talking about Mental Health on Twitter

Post 6

To research and gain insight into how mental health is portrayed, discussed and talked about on the social media platform, a web scraping exercise was undertaken on the Twitter website. It reaped interesting results and observations as Twitter is used by people with different objectives. These include individuals who choose to post on twitter their personal stories or comments about ongoing social issues, some celebrities or ‘officials’ choose it to portray only the positive sides of themselves or some as a way of simply using it as an easy network to connect with people and follow (or sometimes cyber stalk) other people.

To start with a general idea first, my twitter rule using the google sheets web scraper was the words ‘mental health’ written in English, anywhere in the world. The results were immense – up to 10,000 before I had to stop incase the page crashed. Despite the overflow, I did find some interesting posts and how mental health is talked about.

Screen Shot 2016-09-04 at 4.22.39 pm.png

Screen Shot 2016-09-04 at 5.09.33 pm.png

Screen Shot 2016-09-04 at 4.57.10 pm.png

Screen Shot 2016-09-04 at 5.10.30 pm.png

Some were supportive organisations or individuals posting links, photos and other sources to help raise awareness of mental health which then people chose to retweet or comment on. These friendly and approachable tweets were easily found but the tweets I found most interesting were personal, raw and upfront tweets from individuals expressing themselves and their experience or opinion about mental health. To find these more personal stories, I added a rule to find tweets including the hashtag #mentalhealth.

*These posts which were possible to see as their account was on ‘public’ allowing anyone to be able to see what they post.

Screen Shot 2016-09-04 at 5.23.25 pm

Screen Shot 2016-09-04 at 5.23.34 pm

Screen Shot 2016-09-04 at 5.26.13 pm.png

Screen Shot 2016-09-04 at 5.38.26 pm

Going through these findings, it was noticeable that there were much more ‘organisations’ or ‘support pages’ posting links and photos than individuals who comment on their opinion or share their experience with the mental health issue which still highlights the big possibility of the stigma and taboo nature of mental health.

Written by Helen Chang

Post 6: #Researchishard

Post 6: Scraping the web for data
Christine Ye

Twitter Scraping

Albeit not being a Twitter user at all, my opinion of the social media platform is that it is an easy and convenient way for an individual to comment on things going on in the world, express thoughts, personal opinions and even go on a little rant #justbecause. The hashtag system allows people to search for specific topics of relevance, making it a great tool for researching exactly what the general consensus or different opinions of a particular topic are, which is what I tried to find out through my Twitter scrape (yay for freedom of speech on the internet). However, if you’re looking for the opinions of every day individuals in regards to a social issue like housing affordability, the search may turn out to be a lot harder than anticipated.

The Twitter Archiver add-on on Google Sheets was the first tool I used for web scraping, as it seemed like a quick and easy way to filter and collect tweets on the topic of housing affordability – it turned out to be the total opposite. The process of working out the ideal combination of words and hashtags for the search tool was a time consuming process and with my experience I either found 9,876,543,210 tweets by politicians, real estate organisations, online news accounts or individuals posting on behalf of organisations or I found 2 tweets which were irrelevant to the subjective individual opinions I wanted to find. Adding the locational filter of Sydney or Australia also didn’t do much to bring out the tweets I wanted to find on the housing situation within a local or national context, and adding in the filter #generationy also didn’t evoke any meaningful responses.

Using the Twitter Archiver tool on Google Sheets… gained minimal insight due to search rule difficulties.

My second plan of attack was to use the advanced search tool on Twitter itself, and after inputting the same combinations of search limitations as with the Twitter Archiver and receiving the same results I realised… they’re basically the same thing #twitternoob.

On searching hashtags such as ‘housingaffordability’ or ‘housingcrisis’, the results that were shown consisted mainly of politicians and real estate organisations; this was most likely because of the more technical terminology used to limit the search results, as your regular individual would use more colloquial, humorous or long-sentence hashtags in their posts such as #letmewinthelotterysoicanbuyahouse #ijustmadethathashtagup. As said long hashtags are too tricky to pinpoint and search for, I used more neutral terms such as just the combination of ‘house’ and ‘money’ in the word search section, also filtering the tweets to those near Sydney  – this yielded some interesting individual responses.

Some hinted at their ideal housing location/life situation:

Some considered the amount of money they’ve spent or are going to spend on other priorities:

Some talked hypothetical situations #pleaseletmewinthelottery:

Some talked about how hard it is to save for a house in another way:

And Mr C&D raised an important question:

It was interesting to see the tweet above having a relation to some of the articles I read during the first week of this subject, on the aspect of the housing affordability issue being caused by empty nesters refusing to downsize or move out of their homes close to the main cities and into more rural areas. The post puts into perspective the more emotional aspect of housing in terms of experiences, habits and time put into the home that can’t have a price tag put onto it; is there another way to make people more willing to move out of their long-term homes?

Five Point Summary

While the Twitter scrape wasn’t as comprehensive as I would have liked to be, it did turn out to be an interesting method of research.  To summarise five of the more significant findings of my experience, I found that:

  1. In trying to scrape a social media platform such as Twitter where users are a combination of both organisations and individuals, specific terminology may yield only results by organisations or individuals with a strong voice or opinion on the issue that represent organisations with particular motives.
  2. Due to the change and overuse in how a hashtag is commonly used nowadays, it deemed itself to be a hard way to limit Twitter searches if you’re looking for random individual opinions.
  3. For the above two points, using neutral terms that are more common in every day speech in the word search filter will yield more tweets by every day individuals.
  4. In terms of tweets by every day individuals, not many talked about the issue of housing affordability in a direct manner. If a sarcastic or humorous tone of voice was used, it expressed awareness of the issue by the individual however also a sense of helplessness for the situation.
  5. Empathy is a powerful way of giving reason to another individual or group’s actions or thought processes. I feel that because housing affordability isn’t seen to be a very personal social issue, people are quick to make assumptions and quick to point the finger at specific groups. If empathy is used, it could help to relieve the emotional burden or discomfort of some – this could be a good possibility for a design intervention.