Social Media popularity has found its application in hazards monitoring. Please, read the paper published at >>location<<. If you suffer from lack of time there is a very short description of the paper. Your task is to shortly answer the questions:
Q1. Do you think that application of LBSN in hazards monitoring make sense?
Give your opinion (3 sentences only).
Q2. Do you think that Twitter provides accurate and useful spatio-temporal information?
Q3. What are the scientific areas of IT application in social media monitoring?
Paper description
The Internet phenomenon finds a new application every day. Since the social media appeared people still wonders how to use it in a new, an original way. We can share photos, movies through the Facebook or just a short text through the Twitter, etc. The paper shows haw the Location Based Social Networks can be used in hazards monitoring. The article consist of 4 main parts.
Introduction:
The most important aspect of the chapter is that authors characterised a few notations. The firs one is VGI (Volunteered Geographic Information) that means "user generated contents having a geographical dimension". Location Based Social Network (LBSN) are a very important part of VGI as it has a location functionalities combined with SM functionalities, so it plays an important role in SDIs (Spatial Data Infrastructure) as well.
Previous works review:
In the second part Twitter`s main features (messages syntax - hash tags) as a representative of "micro-blogging" Social Network Services are described. Moreover, literature review on Twitter, its application in crisis management, IT tools (API: Application Programming Interface, PHP Data Mining Web Crawling Scripts) that helped in harvesting spatio-temporal information are undermentioned too.
Case Study - the Marseille Fire:
Spatio - temporal aspects of the forest fire (date, time, location) were described in here. Authors underlined, that they had selected Marseille case study because of the low Twitter usage, the fire took place near the area with thousands of citizens - so it was expected a large number of tweets. Four hypothesis for verification were set. 346 tweets were selected through the API - 1 hour before fire started and 5 after it was put out.
Results and discussion:
In the aspect of temporal dynamics "Chronology of the Marseille Fire, number of related tweets per hour and selected tweets’ contents" were prepared (analysis 1).
In the aspect of spatial dynamics "Hectares of burnt area reported in tweets over time"
and "Location, frequency and time of the first citation of place names cited in tweets, and estimated total burnt area" were compiled (analysis 2).
There's a problem, because the linked file cannot be accessed. But I think I found the paper, it is available on Research Gate:
ReplyDelete"OMG, from here, I can see the flames!: a use case of mining Location Based Social Networks to acquire spatiotemporal data on forest fires"
Bertrand De Longueville, Robin S. Smith, Gianluca Luraschi
Wiktor, thank you very much for your attention. I set access to the paper for all English seminar members.
ReplyDeleteQ1. Do you think that application of LBSN in hazards monitoring make sense?
ReplyDeleteGive your opinion (3 sentences only).
Well I like vastly emerging technologies but at beginning of analysis I always search for vulnerabilities and unfortunately there are. But still these technology must upper someday. I think about it as a first version of software framework for mechanism of collecting data from new emerging passive sensors that can broadcast alarm messages when are triggered. Unfortunately I can't find this article about that sensors. I know that this system sound like early Skynet from terminator movie but for now most of machines are not developed to lie or sabotage current systems.
Q2. Do you think that Twitter provides accurate and useful spatio-temporal information?
I think rather not. At beginning I believe technology work fine but sooner or later will upper people similar to those who order pizza on the 112 telephone number.
Q3. What are the scientific areas of IT application in social media monitoring?
I watch progress of who internet connect people, what profession it represent and how its evolving. One of interesting area that social media could monitor would be political empathies that could boost changes in political and law system. My dream is to create contentious virtual democracy but it require truly interdisciplinary knowledge.
Piotr, thank you for your interesting point of view. Having read your comments I would like to give my opinion.
DeleteQ2: I agree with your up to a point. There are two areas of the question. The first one is connected with geographical location of the tweets (coordinates) whereas the second one refers to their content. I agree that tweet`s content is mostly comparable to a non-emergency calls but their coordinates spring from a device location, so they should be accurate.
Application of Social Media has been spreading like wildfire but I haven`t heard about creating contentious virtual democracy. It is said that operation on Social Media contents focus on text processing. Do you think to achieve the idea you dream about, the undermentioned scientific areas can be applied: preprocessing ("transforming input data such as text for use with machine learning algorithms"), classification ("Identifying to which category an object belongs too"), clustering ("Automatic grouping of similar objects into sets"), regression ("Predicting a continuous-valued attribute associated with an object")? [Citation from: http://scikit-learn.org/stable/]
DeleteQ1. Do you think that application of LBSN in hazards monitoring make sense?
ReplyDeleteGive your opinion (3 sentences only).
As much as we would like things that are already popular to fulfill some other functions, usually they let us down in this area. The authors of the article pointed out that in the initial phase of the event there were innaccuracies between data provided by Twitter and what actually happened (due to media hype). Perhaps if we could filter out such cases the method could perform accurate enough.
Q2. Do you think that Twitter provides accurate and useful spatio-temporal information?
I think same as Piotr, Twitter wasn't created for that specific purpose. Also I laughed at the pizza case, as I read about it some time ago. :)
Q3. What are the scientific areas of IT application in social media monitoring?
Social media monitoring is quiet ofently used to predict trends. Also there is this article that states that predicting the future of how well a product will do can be easily done through social media monitoring and that companies already use such tools (source).
Q2. Take a look at Piter`s Q2 answer.
DeleteQ1. I agree with you that in this particular case there were incompatibilities between data provided by Twitter, media and what happened in fact.
DeleteYou mentioned about filtering mechanism. I agree with you entirely, that proper filtering methods could be the key to the success. So, some methods like machine learning, Python text processing (NLTK, http://www.nltk.org/), scikit learning (http://scikit-learn.org/stable/ ), statistical language processing NLP and social media mining are applied.
Michail, you mentioned in answer to Q3 about predictions. I underlined some areas of text processing - look at comment for Piotr (19 April 2015 at 14:58). Have you ever heard about examples of predicting algorithms? I`ve heard for example about the isotonic regression which "finds a non-decreasing approximation of a function while minimizing the mean squared error on the training data" [Source: scikit-learn.org].
DeleteQ1. Do you think that application of LBSN in hazards monitoring make sense?
ReplyDeleteGive your opinion (3 sentences only).
yes, we can use this technologie, but we must analysis the information before . the social media are one of the most recent and popular telecommunication technologies. Despite that such websites were built to be used between individuals, their use has extended to criminal activity by exchanging information on assaulting individuals or upsetting public peace, in addition to incitement to commit acts of terrorism and riot. Unfortunately, the number of social networks users spreading destructive ideas affecting our society in our current time is increasing.
Monem, Thank you very much for your opinion. You are right that before we use Social Media as a source of data, it is necessary to analyse and select information first, as they spined out of control (planned functionalities). As my point of view referred to a data analysis, please take a look at my answer to Michail (Q1).
DeleteQ2. Do you think that Twitter provides accurate and useful spatio-temporal information?
ReplyDeleteTwitter can provides a lot of information for spatio-temporal, but the problem not valid,so for me this technology for lost the time only .
Thank you for your opinion. But why do you think so, I mean "... so for me this technology for lost the time only"?
DeleteQ3. What are the scientific areas of IT application in social media monitoring?
ReplyDeleteSocial media monitoring goes beyond customer care, however. You can use it to track trends, monitor public attitude toward your company and perform market research.
Q3. I would like to ask you the same question as I did during conversation with Michail (19 April 2015 at 15:09) as you mention about track trends.
DeleteHave you ever heard about examples of predicting algorithms? Do you know any examples?
Q1. Do you think that application of LBSN in hazards monitoring make sense?
ReplyDeleteGive your opinion (3 sentences only).
It does make sense. In case of an emergency other people in the area who participate in those networks can be notified more quickly of a danger. That may help to reduce the chance of a panic outbreak when nobody knows what exactly is going on.
Q2. Do you think that Twitter provides accurate and useful spatio-temporal information?
One would think that during a true disaster of big enough proportions, nobody within reach will have the time or even the possibility to send a tweet. Then again, I remember a funny picture of a man on a segway using his smartphone whilst escaping a wildfire (here it is: http://newsimg.bbc.co.uk/media/images/45211000/jpg/_45211426_segway_getty466b.jpg).
Either way this source of information will be lacking in accuracy. The Marseille forest fire case study from the article may have yielded good results but analysis such a disaster in hindsight is much different to real time monitoring.
Q3. What are the scientific areas of IT application in social media monitoring?
I'm sure there are many ranging from political science (for example monitoring riots) to gathering data about the spread of hypothetical pandemics. There's also the question of security. Sometimes damage control requires keeping critical information out of social media (such as in a hostage situation).
Q2. I`m solid with you. If an emergency appears nobody will tweet, though you found a very funny pic :). For me SM can be applied for hazards predicting that are built up for a long time, like floods, terrorist attack, etc.
DeleteQ1. Giving my opinion to your answer, I agree that SM are one of the many information channel (TV, SMS, radio) during crisis management, like floods, heavy snowfalls, rains or high winds. But we agree that SM are not the only one.
DeleteQ1. Do you think that application of LBSN in hazards monitoring make sense?
ReplyDeleteGive your opinion (3 sentences only).
It might be good additional 2nd grade tool in example for analysing post factum information like people’s emotions, but it won’t ever offer equal quality like simple alarm phone. What would you do in case of direct danger - post a message on social platform or call police/ fire brigade? I think you would make a phone call and then eventually start typing, so phone information would arrive much quicker and proper services wouldn’t need duplication in the Tweeter post.
Q2. Do you think that Twitter provides accurate and useful spatio-temporal information?
I think, those information won’t be ever accurate. Platforms like Twitter will always provide information limited to its internal users. Lot of people won’t ever open account on platforms like that i.e. because of the privacy policy reasons or just because they don’t like to publish them self. So Twitter statistics will always miss a huge part of reality.
Q3. What are the scientific areas of IT application in social media monitoring?
I think that future direction is in word’s stream analysis and accumulation of additional meanings from what people are typing, including statistical outputs. However even best algorithms which are analysing human’s semantics still fail, please try to translate any complex sentence in Goole’s or Bing’s language tools.
This comment has been removed by the author.
DeleteAlbert, Thank you very much for your opinion. I`m sorry I didnt`t answer but I was 10.000 meters above the ground during my business trip while you published your comments, so I wasn`t able physically to give my opinion. I do it right now.
DeleteQ1. I agree with you that emergency phone number can`t be replaced by SM. That is out of question. Post factum information analysis is quite interesting and safe approach. For example we could searching for and selecting information connected with terrorism (see: comments for Wiktor at 19 April 2015 at 15:19).
Q2. I am afraid you are right. Majority of people don`t share with their location cause of the privacy policy reason.
Q1. Do you think that application of LBSN in hazards monitoring make sense?
ReplyDeleteGive your opinion (3 sentences only).
IMHO Yes. In case of fire or anything like that, an approximate location from LBSN would help to locate and approach in time. I'm curious about the other events (we need something better than an approximate location) and a possible lag on the network. Maybe in time, the social networks may prioritize the delivery of some messages.
Q2. Do you think that Twitter provides accurate and useful spatio-temporal information? Depends to the information.
Q3. What are the scientific areas of IT application in social media monitoring?
I don't have much opinion on this. Can I say, umm. profiling ?
It was a nice reading. Thanks.
Albert, Thank you very much for your opinion. I`m sorry I didnt`t answer but I was 10.000 meters above the ground during my business trip while you published your comments, so I wasn`t able physically to give my opinion. I do it right now.
DeleteQ1. In my opinion, as I commented Albert`s line emergency, phone number can`t be replaced by SM, but you pointed out very interesting area of application connected with location approximation.
Thanks for your opinions.
Q1: LBSN is certainly a useful tool for analyzing the levels of emotions of the users with regards to a certain place or event, that's why marketers and city planners for example can gain valuable information on what place on the map is popular or needs action. However with regards to hazardous events, there is a need for accuracy and in-time reaction - the tweets are emotional and therefore subjective and may not give the correct picture - the monitoring system would need to take this into account and provide the users with a set of instructions on how to conceive the message.
ReplyDeleteQ2. I am not sure about that. It all depends on how it is analized
Q3. Certainly the city planning, as mentioned before, and spatial economy. Scraping social media provides a powerful image of what may be the social and development need of a city - there are not enough bike routes? A place on the city map is particularly popular? Why? And how can this be replicated in different places? And more generally - does this kind of popularity depend on specific spatio-temporal features of the place, and so on.
Dominika, Thank you very much for your opinion. I`m sorry I didnt`t answer your comment on time. I do it shortly right now.
DeleteYou are right. During emergency location accuracy, information without emotions and in-time reaction are highly required. On the other hand emergency services should monitor all information channel, so the package or set of instruction for citizens "how to do it" are welcomed.
Thank you very much.
1. Do you think that application of LBSN in hazards monitoring make sense?
ReplyDeleteI think that every additional source of information in hazards monitoring does make sense, and this particular source can be very useful. Please note that this paper describes a fire event from 2009 - smartphones were much less popular back then (iPhone was released in 2007, Android in 2008, global smartphone sales in 2014 were around 7 times higher in 2014 than in 2009 and they’re still growing). So with time there will be more and more users providing useful information.
2. Do you think that Twitter provides accurate and useful spatio-temporal information?
It’s hard for me to say as I don’t use Twitter, but from what I see in their ‘Developer Geo Guidelines’, users can agree to publish their exact location. So probably yes - Twitter (and probably other similar social sites) can provide quite accurate information. I think that the growing number of smartphone users will also result in higher ‘density’ of information, so probably this information can become very useful and accurate.
3. What are the scientific areas of IT application in social media monitoring?
I think that social media can be a mine of information for social science and crisis communication.
Q1. You are right.The popularity of mobile devices has been growing since 2007. I am wondering how many information in the aspect of hazards have been in recent years and if it can be useful.
DeleteQ3. My opinion: see comment 19 April 2015 at 14:58 as an answer to Piter. Scientific areas that can be applied and developed: preprocessing, classification, clustering, regression, so all connected with machine learning and text processing and social media mining [see: http://scikit-learn.org/stable/]
Thanks for a reply.
DeleteQ1: Unfortunately I have no idea, although I'm sure that the number of people using social media on their mobile devices is growing. You could probably browse papers citing your paper to find out if there's been some progress in this area.
Q3: Thanks, I was trying to come up with something new but there wasn't much left to be added to the list :)