Original Post at Business Fight Poverty – http://community.businessfightspoverty.org/profiles/blogs/personalization-in-technologies-for-developing-regions
Heterogeneity is a natural aspect of life. From macro to micro, everything involves variation in data at a great level. As much important it is to find common patterns among everything, it is equally important to value the specific traits of every category. Personalization addresses this particular objective and helps develop solutions targeting specific segments of a bigger group. Personalization can be at two levels. One personalization can be done at the level of technology. This means that technology and underlying protocols have been designed specifically for a particular environment. The other level of personalization is seen at the level of content, where a common platform provides data in a personalized manner for dedicated segments.
Implementing technical solutions in developing regions face challenges similar to these two levels of personalization. Barriers like illiteracy and poverty force the content to be highly personalized for certain communities. And poor networks and large gaps in economic distribution create a need for different technologies and platforms.
In this post, I’d like to discuss the needs and potential applications of personalization in developing regions. Applications in developing regions generally target global development challenges like education, nutrition, poverty, health, etc. In our studies, we have focused on the problem in agriculture development. In India, agriculture is primary rural occupation and most of the agriculture workers come from remote villages. These are areas with poor infrastructure, little or no education, and unstable electricity availability. There is a large amount of assistance that can be given to people in these areas with the help of information and communication technology, but it has to be specially designed for dedicated cause. Personalization therefore is required at a very basic level here and entire system design has to be built in a way that it is able to adapt itself to these conditions. Personalization helps modify existing technologies for specific application and avoid the cost of developing new solutions from scratch.
First I’d discuss different constraints in developing regions that act as parameters in system design. Then I’ll describe the four particular areas where personalization needs to be applied, followed by a discussion on personalization integrated development.
Parameters for Personalization in Developing Regions
Systems being designed to help people in developing areas with their livelihood or occupational activities need to tackle these constraints. These constraints can themselves be used as parameters in design.
Illiteracy is the biggest constraint faced by users in developing regions from using technology. Even if today mobile phone has penetrated into these regions, users are still not able to use all features of even basic phones. Most of the users are not able to perform tasks like sending SMS from phone or setting up a ringtone. Moreover, inability to read text makes the use of phone book impossible for many users. This creates the need to design specific interfaces for people from these regions.
Networks are at the base of most of our technical solutions. Broadband internet is not available everywhere in these areas. Cellular data networks are also very unstable in these regions as inter-tower distance can be large at many places. Systems using internet data here need to collect data on network availability and buffer it if network gets unavailable. Standard systems with respective time outs cannot work well in this environment.
Electricity availability is a very big concern in these areas. There are long power cuts throughout the day. Relying on direct power connected systems is impossible. Generally, battery operated systems or ones with backup power ultimately work in these regions.
People in these regions are mostly superstitious and keep very strong hold on traditional beliefs. Their perspective towards technology can be very different than expected. It is not easy to make them adapt to technical systems and make them a part of their routine lives.
These problems and many more small issues act as barriers in developing technology for these regions. Personalization can focus on these issues and once a system identifies them, it must be able to change that particular aspect according to the environment. Personalization at this level becomes rather challenging as language support and design constraints can vary up to a large extent among different development regions.
Areas of Personalization in Developing Regions
Personalization in Interface Design
Interface to a technical solution is what mainly matters for a user. Interfaces can be on console, hardware devices, software UI, mobile phone interface, etc. In the process of designing assistive technologies in developing regions, most common interface is mobile phone. At present, it is one of the best available medium to design ICT solutions for people in developing regions. Another medium now being used in rural areas is laptop computers. GUI operating system equipped laptop computers can allow a lot more technical solutions to be made available for the users.
Interface personalization in these systems is mostly based on educational parameters. Mobile application or mobile software interfaces can be designed to suit low-literate and illiterate users. Creating mobile apps to suit design needs is easier than OS design. An app can have both general version and a specific user group version. But creating such different spins of the complete OS is not as easily applicable. A good design can be one where a particular version of mobile OS can itself accommodate some design changes as user base changes. This opens up an area for research where multiple designs can be placed within a single OS with minimalistic differences.
Similarly, interface in desktop applications or web applications need to be designed specific to user base. Web applications before HTML5 rely heavily on additional software. This software needs to be updated continuously. Now in these regions, neither the network is consistent enough to ensure regular updates, nor are the users much aware of advantages of updates. Usually even at places with internet supply, internet plans come with small downloadable data limits at affordable prices. Users in such case avoid updates to prevent data download. Therefore, building web solutions including multimedia should be mostly independent of additional software or browser constraint. At a gap of reasonable amount of time, these software and browser can be updated, but design of web applications should be such that they can run effectively in almost every scenario. For eg., HTML 5 doesn’t need any additional software support for video or audio. If open source format is used for data, then almost every browser will run it perfectly. Therefore, with web applications targeting rural users, personalized design has to be developed for proper effectiveness. Design constraints based on literacy apply here as well. Interface of web applications must be made as much comfortable as possible for these users so that they can get some advantage out of it.
Personalization in Network Protocols
Network is critical to almost every application directed towards development. Information and knowledge systems are being designed to address needs of these regions. Guidance systems being designed are mostly depending on internet sources. This makes network personalization a very important requirement. Applications under standard conditions run on standard network protocols. This means that there is continuous network availability at a consistent speed. Default connection time outs and waiting times in applications generally accommodate well with the network and no issues are observed. Same application can create issues when used in such regions because network won’t be consistent. Even when available, network speed can be a problem. Protocols at network and transport layer need to be modified for such applications.
Systems can be designed where an underlying network framework monitors network statistics and modifies its functionality. For eg., Buffering of packets and temporarily storing it within framework is one way of gathering network data in inconsistent network environments. Many other protocols exist and are being developed which manage data in such networks. In order to bring rural technologies into real effect, this understanding of network personalization is very important.
Personalization based on Platform
Platform based personalization essentially means changing the format of application as per the end user platform. Some application developed for desktop systems might have relevance in rural needs. Desktop applications can only be used in rural colonies through internet café or laptops that are still very scarce. Ideal platform of applications at present in these areas is mobile phone. In the previous three years, there has been a growth in new segment of Smartphone which are low cost and low on hardware configurations, but affordable up to a great extent. With more rise expected in the distribution of these low cost Smartphone in rural areas, a large number of applications can be ported to mobile OS platform.
This particular personalization can be identified through multiple surveys and market study. These combined with location can help suggest the application developers, what platform to target first. There are areas where internet has reached and is being used regularly. Making web clients for applications can be effectively used in such areas. Platform personalization has proven to be useful in general as well. On desktops or laptops, usage of mail has mostly been observed through web clients. Apart from corporate users, majority people use web clients for their mail service. When shifted to mobile OS, these same users have preferred mail client applications over web mail. Web clients can still be used on mobile, but with change of platform, user’s preferences change. These are educated users who can comfortable handle any kind of application on different platforms, but a good understanding of their preferences help increase utility of technology and hence make it effective. Using this approach in developing regions seems far more relevant as anyhow these are environments with multiple constraints. Finding best suitable platform of deployment is very important for developing usable and effective systems.
Personalization in Content
Content personalization is among the most used kind in present day systems. Social networks, advertisements, search engines, etc. are particularly adding personalization in content. This in fact has been a vital point in any system’s success. People tend to look for the “recommended” section in most of the applications while making a choice or decision. It holds equivalent importance in technologies for developing regions as well.
Applications targeting challenges in developing regions need to be very simple and precise. In our study and experience, we have observed that multiple features in single application are confusing and irritating for users in these regions. Information being provided by a service should address particular needs of the user. Otherwise they quickly discard it without making attempts to adapt. As mentioned before, adaptability and acceptance for technology is very critical in these regions. If they develop any kind of negativity or fear with some application, then making it useful for them becomes very difficult. For eg,, we have developed a news summarization and distribution system for agriculture community. Our users include educated agriculturists, traders etc., and lesser educated agriculture workers like farmers, local market operators, etc. Information being sent to the lesser educated subscribers here has to highly filtered and made as much comfortable for reading as possible. We categorize our content, translate it to local languages and send them in their preferred format of SMS. Similarly other applications for such users also have to be highly personalized in terms of content.
Personalization integrated Development
The four areas of personalization discovered specifically for developing regions act as a base for developing applications in these conditions. Our focus has to be on finding solutions to the large number of problems faced by people in developing regions. It is true that technology can help provide support in many ways and can actually act as a tool for development workers as well. But it has already been observed that using standard existing technologies cannot work optimally. If our systems have to break the literacy and economic barriers, then they will have to be dedicated designs built by understanding the user base. Application development integrated with personalization in the areas mentioned in previous section can be a great help the users and developers. Instead of developing entire different versions of software, we can add as much personalization as much possible based on respective parameters.
Our idea with integrated development aims this particular need oriented development which can help convert standard software into multi dimensional software with a branch of it adapting itself to the developing regions. Including data elements from social studies, application domain datasets, demographics, etc. can help make such systems for developing regions and ultimately help in global development.
I’ve tried to relate the concept of personalization to a very specific area of technology design and development for developing countries. Development objectives discussed here co-relate social problems with technical specifications and help identify focus points for researchers and developers working in the area of ICTD. This development would include a large amount of statistical and social data across multiple dimensions. But this is one way through which we can achieve the ultimate goal of global development. Information and communication technologies for developing regions often face issues usually unknown to computer science community. Through this post, I just want to highlight few basic constraints that highly control applicability of ICT solutions in developing regions. If we can take care of these small areas while developing solutions, we might be able to help the developing world better than today!