Jumped here from twitter to support this idea of Urdu chatbot 
.Best of luck whoever came up with this idea and is developing it for the community 
6 Likes
Hi Nafezukh, thank you for finding your way here and taking the time to write your view on this. I can’t wait to have the finished product so that you and others can benefit from this tool.
4 Likes
I am from Pakistan and its important that the community have the access to chatbot. i have read and will say I am agree what @asadus16 said that it should be added somewhere where people can interact with it & get used to it. Hindi Urdu community need these types of project building because it will help me and all others.
4 Likes
Hi Ali, thank you for you kind words! Much appreciated 
there chatbots will be linked into the community telegrams where I have already added the welcomes messages and security bots. this should hopefully make them easily accessible if people forget where they can find them. I will update the communities as I build these as I ask for Beta Testers for the official release.
4 Likes
Hi @GRTDataHub ,
I have a few questions to which I can not find the answer too, please ignore them if you have already answered them.
-
Was the file-structure of your chatbot changed to a system where all graph advocates could potentially upload the Q&A of their own language and you would only have to merge?
-
What is meant with
Labour, Development and Continuous Improvements? What exactly are you getting a grant for if this is your proposed work? Why does it cost exactly $2000 per new language? -
Will the cost of your next application go up in linear fashion?
-
Is someone else able to plug and play the chatbot if they would like? Example: Your server goes down and I would like to host your chatbot, is that possible?
Where I see a lot of positivity coming from multiple communities, I am currently voting no on the proposal. My uncertainties from last time are not gone and the amount of funding requested keeps increasing. In my opinion, this does not seem sustainable in the long run.
I will gladly change my vote if the answers to my question change my mind, I am looking forward to your answers.
4 Likes
Hi @Colson,
Thanks for your questions and I will answer them and hopefully clarity some points and get you to reconsider your vote.
Please visit the app and you will find at the bottom of ALL chatbots is a Translator Help Page button, within this page you will find the information regarding how advocates / community members can improve and contribute towards the continuous improvement of each chatbot. There are instructions provided on how to access the code via the open-source GitHub. How to create a pull-request to collaborate the code with others. There is also a link provided to the discussions forum on that Github regarding new content and translations where anyone can contribute if they don’t feel confident in making the pull-request. I will then add this myself. At the bottom of that page you will also find 2 buttons, 1 will download the chatbot content in Microsoft Word and the other in a Text file. These can be used for translations of new chatbots if and when required.
Link for GRTDataHub GitHub, GitHub - GRTDataHub/GRTDataHubDevelopment: GRTDataHub App for Data Anlaytics and Educational Chatbots built for The Graph Communities. , please scroll down and read the README.md file for more context and instructions.
Please on the end of my last grant I was asked to update the committee on the above so I also provided it there here, Grant Proposal: French (Français) & Turkish (Türkçe) Chatbots and additional app features - GRTDataHub - #34 by GRTDataHub .
Labour, Development and Continuous improvement to myself is self-explanatory however I will elaborate and add context. All of the above are the time, effort, community engagement and working with translators, research, chatbot continuous improvements, upkeep. It is the work being carried out to develop the chatbots into the tool and final product ready to be used and beneficial to the community it is intended for. The time it takes me to make each chatbot from start to finish can differ. For example, as per this grant the Urdu and Hindi chatbots will take me almost 1.5x as long to make as previously due to the complexity of the language and cursor logic when updating the code, links, documentation and logic due to the sheer amount of work provided. On that note, I spend around a month of each chatbot, between 3-4 hours per night. Over the course of a month this adds up to above 80-90 hours give or take a few days off. If you have asked a translator who has worked with me about the amount of work provided for them, and then that needs to be dissected between 5-20 times PER chatbot answer provider into the code and links put in the correct places with no issues, no bugs which breaks the chatbot along with unique links and content from that community. You start to get an idea of the amount of work that is required here. I have been asked on previous grants to provide an hourly figure for the work carried out but this isn’t feasible as it would only be around £18-22 per hour for the work taken to build each chatbot. I price these per chatbot and not per time due to that reason. I love building these and I learn something new each time I do, personal development and working amongst new communities making new friendships and connections. Even though the Urdu and Hindi chatbots will take longer from what I have already build and experienced, this is not reflected within the grant amount per chatbot.
On your point 3 regarding the next application go up in a linear fashion. I don’t understand your question here or maybe you have not read the grant information regarding the previous grants approved for the work carried out. You have previously commented in the community call regard the amounts increasing per grant which I addressed at the time as you also mentioned it not being sustainable in the long run and this was explained on the community call. My apologies if this has been forgotten but I would like to clear this up to prevent this reoccurring question. It was due to 1 chatbot, then 2 chatbots and now 3 chatbots being build per grant. Therefoe $2,000 X the chatbots may answer this recurring question. Please scroll up to see this information;
The Graph - GRTDataHub App Analytics - Wave 6 = $2,000
Graph AdvocatesDAO
English chatbot = $2,000
German and Spanish Chatbot = $4,000
French and Turkish Chatbot = $4,000
This is no different to a grant being submitted for hosting 1 event. Then hosting said event the next time in 2 places incurring double the cost and labour etc. However, this is development, guides, tutorials, educational content and community tooling, not a hosted event. I hope that analogy makes sense.
Another note, regarding the sustainability in the long run, I believe this will be the last grant submitted by myself for community chatbots. I believe the tool is a great asset to each community and the feedback I have got has been unbelievable. Along with the excitement of new communities waiting in anticipation for this tool as per the comments above and via Twitter and their community telegram chats proves this. However, I would like to step back from making these chatbots after these 3 and concentrate on the continuous improvement on the current ones (as per the grants) to ensure they are as good as possible before the influx of new GRT holders during the next bull run. I believe each one will be used to educate, onboard and grow the graph community and create advocates, which is why I want to take a step back from developing these and create a “Quiz me” feature that will help test, prepare and aid community members to smoothy transition into potential advocates using the educational content within these chatbots.
To answer point 4 regarding if anyone is able to “plug and play”, which the GitHub being open-source this means anyone at any point can fork this repo and also build as long as they know how to code in React.js. Being a novice coder I am sure a lot of developers will be able to do this with ease. As mentioned in the first paragraph, please see the forum post on the previous grant or README.md file regarding the License being Apache-2.0. Meaning it is available for anyone to use. (This can be read within the License file within the GitHub repo)
I believe this grant ticks all the boxes for the product / tools, its use and value within the community. Increasing community awareness, engagement, understanding or participating in The Graph or Web3 ecosystem. Could you provide context being the No vote regarding the above as per the community grant documentation please? I would love to hear your thoughts on my above replies as I hope I have clarifies a few points and answered your questions in great detail. And hopefully you reconsider your vote after reading this. 
Kind Regards,
Liam | GRTDataHub
7 Likes
Hi Kyle,
I have received a message from 2 community members from the Urdu and Hindi communities who have commented their thoughts and support this grant. However, due to their rep they are unable to vote. What are their actions to be taken and will this be taken into consideration? I am hoping these tools will help to onboard potential advocates from these communities in the future, however there is not many as of yet.
Will this be taken into consideration given the support shown in the comments, via Twitter and their Telegram communities?
2 Likes
Thanks Liam! I love what you have done with Chatbot and you’re work around the whole community.
Before issuing more funds toward the Chatbot, I’d feel more comfortable once we have some more usage statistics. So far there has been $12,000 already rewarded for the development and translation efforts for chatbot. Is the $6,000 that is being requested the best next step for chatbot? Would it make sense to consider funds toward implementation in a greater sphere?
For now I have voted no, but happy to change that to a yes with some additional information 
Derek
3 Likes
Hi Derek,
Thanks for the question and thanks for your kind words regarding my work so far.
I am still in the process of adding the data analytics however I am unsure how this will add weight to the grant, as the work carried out, there is often uptake in site visits around the release of a new from the country in which it was intended. Which then carries on to get site visits thereafter.
So an example is from the last 7 days, the reoccurring users from the countries that the chatbot has been made for already.
The total number of requests from the past 7 days below…
For example if an event like, delegators able to get rewards on Arbitrum and want a translated guide on this. There could be an increase in site visits that month and the community then keeps visiting the app.
Could you please elaborate on your question? I will try to answer it however, are you implying that the $6,000 wouldn’t not be useful to build 3 more chatbots to help educate and onboard 3 more communities of enthusiastic GRT holders to help grow their community with provide them with translated content, documentation and tutorials which will aid them in guiding future community members and promote advocacy.
Could you please explain what you mean by using these funds to instead implement them in a greater sphere because I don’t know what this means?
Sorry to not have answers your questions just yet.
Kind Regards,
Liam
1 Like
Great! I saw this on an earlier post. But I think I’m better understanding it now.
Since I am not connected in the Urdu, Hindi and Ukrainian communities I don’t have a way to measure what type of existing userbase this would service. Are there telegram communities focused on these languages? If so, how many community members are in them (thus giving us a better idea of the benefit of translating it to these languages).
Are there other implementations outside of telegram that would be worth integrating with? For example, I know The Graph discord server has recently implemented AwesomeQA in their discord. Would we get more bang for our buck if there was a potential collaboration between the effort to train AwesomeQA with some of Chatbot’s existing information?
2 Likes
Hi Derek,
Okay that makes sense, I will try and provide the details for you. These communities are fairly new to be set up however I have aided the community admins and set up the security bots, welcome bot and links feature into their community Telegram. These groups are new but we have grown their Twitter account through the admins being active, proactive and also through sharing GRT Daily Statistics, growing the community engagement and followers week by week. I will share the current figures below.
I am also always happy to help any community that would like this, as you can see from above I have helped implement this across many channels. We are working to increase the community members of these communities through the Daily GRT Statistics which does grow the community week on week. I have a group of community admins, community support and newsletter authors. This helps us to easily collaborate and support one another, also, I ensure all of these are within the GRTDataHub app (Graph Advocates Spotlight and Graph central both have their own page where I archive all issues in all languages for future and current community members to be able to easily access). DappLooker dashboard integration within the app and a new feature I am working on is a carousel of the communities within the GRTDataHub with direct links to their socials and chat groups.
Below are the currently members. Together I believe we make the graph ecosystem and groups better by the day by providing support to one another.
I hope this paint a bigger picture for you on the work that I am carrying out and how these chatbots are a part of the whole onboarding and community tooling process to improve new communities.
I have also recently seen the AwesomeQA which has been implemented recently, I believe Colson is the one training this is that correct? He asked to provide this bot with content for training etc. which I said I would provide these questions.
All of these can be accessed within the open-source GitHub within the ENGMessageParser.js which taken the user input and then provides the answer to the input.
However, with these being used to train the Discord chatbot I thought this may be a duplication of the work already carried out to build the English chatbot within the GRTDataHub app. I have always believe the community chats should stay human community led and then maybe a direct link to the chatbots already built and fit for purpose.
4 Likes
Hey @DataNexus ! I agree with liam about the new Pakistan community that have just started to build and is gaining more attention with time as i am doing my best to spread graph awareness by organising educational events, translating text to help my community and by social interactions specifically on twitter(for now). Also you can check my recent event here
https://twitter.com/thegraph_pk/status/1623720007262945282?s=46&t=yfY64eRYHk2Zw1gxZPFLzA
And will be organising one on 3rd March !!
This all is too support the thing that Urdu community is getting stronger with time and as a community lead in Pakistan i believe this chatbot can be a help to our community for sure 
3 Likes
I think this would be awesome. Not sure that something like this falls within the current grant request, but I see a lot of value in expediting the AwesomeQA knowledge base.
Very cool! It’s understandable that the communities would be new, and this is something that aims to help them grow more rapidly.
I have changed my vote to a yes with the data you’ve included
2 Likes
Derek,
Thank you for taking my time to read my essay and sorry if it wasted your time haha. But I am glad I have managed to give you the full picture.
I would be more than happy to help! However, it has also been highlighted that the use of AI would be cool to see within these chatbots for the future. I have not previously explored this avenue but maybe it would be the next logical step (I will seek advice on this when the time comes as to the best way to approach this). Once this major upgrade has happened then that tool will probably be the most powerful, with the use of neural network so the community can teach the bot the more it is used. My apologies if this isn’t correct as I have not yet explored this route however I would love to help in any way I can.
Once again, thanks for your time and in reading my responses to your questions.
1 Like
That’d be awesome! Looking forward to it!
3 Likes
Thank you for your answers Liam!
If I have any more questions later I’ll reach out to you in this forum.
3 Likes
I’m torn about the application, I see that the communities are actively using it, which means it’s helpful. I also know that you are very hardworking and I admire your attitude and #buidl mindset! So here I’m arrogantly forcing even more development from you! 
You know already my opinion about how the translation process of the chatbot should look like. I can’t see that it has changed the way I hoped and kind of expected after your last enthusiastic reaction to the Grant Committee’s suggestions. The language versions simply cost too much ($2000 for dissecting the code + $300-$500 for each translation). My opinion is that with the proper architecture, we should be paying only for the translation.
Another aspect is the continuation/maintenance of the translation for these and previous languages. I imagine there is content added to the chatbot’s questions and answers set all the time and it’s confirmed by the growing number of words needed to be translated in the source file, which grew by about 2000 since November/December. Are the previous translations of the chatbot updated as well?
With that aspect in mind, re-writing the code should allow more time and cost-efficient, continuous translation. So, re-writing the code grows into an even bigger problem with new urgency (for me). Maybe are there some technical teachers who would be able to help with it, as I know it requires even more time and effort, but I believe it is worth all the resources in the long run.
I’m also concerned about the future of the chatbot in the context of the recent hype around e.g. chat GPT… It is probable that it all goes to waste, as AI solutions are getting smarter and more available to everyone in all languages. That’s why I appreciate your willingness to look into AI solutions to experiment and again, in my opinion, you should be more focused on the technical development of the chatbot.
I’m also not entirely sure about the Ukrainian demand for a chatbot, as there are only 11 members of the unofficial community, including 3 people in that group who I know they’re not Ukrainians. I’m wondering why hadn’t you chosen the Russian language for this round of translation instead, as the Russian community is significantly bigger (1,2k people).
So for me, it is YES in the community pool.
NO for sustainability, the efficiency of the translation process and its current cost.
2 Likes
In general, on this point I can add that I personally have some plans to develop the Ukrainian community, so such a bot will be a pretty cool addition and will help introduce new people to The Graph.
About the russian group, at this stage I haven’t contacted them in any way, so there are no members among the DAO who can help manage this group.
But, I will take care of this issue and will be able to help with translation and further work on adding this bot to the russian chat and setting it up.
Hi Julianna,
Thank you for your kind words.
I am unsure what you mean by “you can’t see that it has changed”. Previously I would send the chatbot transcript and all the content that needs translation and the translators would sent it back so I can make all the changes. I planned to update this as I agreed it is not good for the continuous updates for the chatbot. So I did in-fact re-think the whole process and made a lot of improvements.
Please see the comments within this forum for these details however I will also provide these again for you:
Please visit the app and you will find at the bottom of ALL chatbots is a Translator Help Page button, within this page you will find the information regarding how advocates / community members can improve and contribute towards the continuous improvement of each chatbot. There are instructions provided on how to access the code via the open-source GitHub. How to create a pull-request to collaborate the code with others. There is also a link provided to the discussions forum on that Github regarding new content and translations where anyone can contribute if they don’t feel confident in making the pull-request. I will then add this myself. At the bottom of that page you will also find 2 buttons, 1 will download the chatbot content in Microsoft Word and the other in a Text file. These can be used for translations of new chatbots if and when required.
Link for GRTDataHub GitHub, GitHub - GRTDataHub/GRTDataHubDevelopment: GRTDataHub App for Data Anlaytics and Educational Chatbots built for The Graph Communities. 2 , please scroll down and read the README.md file for more context and instructions.
With regards to the continuous improvement of all chatbots, this is done within GitHub. I send out a message “announcement” of new content ready for translation. I add it to the code and comment it out as it hasn’t yet been checked and approved by a translator. I have seen that this hasn’t been done for a few bits “this is a new improvement of mine that is being trialed”, I have started updating it when I have over 200 words to translate rather than each and every time. So it doesn’t waist the time of translators / anyone wanting to update and contribute to the code. As it is all open-source.
I have already re wrote the code, however I am having a few issues with some minor bugs so I have not yet pushed these to the live app as it would stop it from working. Which I don’t want to do, as re writing the code has no effect whatever on the end user or the chatbots usage etc, however I have done it as per the old “suggestions” from the previous grant. However I do not see how the time taken for me to update the code, around 2-3 weeks so far, whilst hitting obstacles that I am now having to de-bug before I can push it to the live app. Will have any knock on effect to the translators? But, I have indeed made steps to resolve this questions of yours.
I am also in the process of re-vamping the theme, (UI and UX) of the app whilst I wait for the next section of translations to be sent through. (Translator had prior commitments to I didn’t want them to rush getting there to me as I don’t want to add any pressure). So I am also working to improve the app as a whole alongside improving the chatbot code (which for me is a slight wasted effort as the end result will be no different to how the chatbots currently operate).
I understand your concern about chatGDT however I will be looking to implement something similar in the near future. I disagree that it will all go to waste, that could be said for anything in the world to do with AI taking over. If we all stopped building then the ecosystem would not he continuously improved so I don’t see the logic there sorry.
The point of the chatbots are to help grow, guide, onboard and bring more people into The Graph ecosystem. Surely targeting the communities with proactive advocates / DAO members that are currently trying to bring in more users and create a tool to do so is the way compared to making a tool for a community that is already thriving? So I apologise but I once again disagree with you there, unless you only measure the success and usefulness of this tool on the number of users from that community visiting it. Which for me, should not be the only focus point here.
I am not going to take the time to explain the sustainability etc please could you scroll up and read these in a previous answer of mine.
I hope this answers your questions.
Kind Regards,
Liam
Thank you Liam, I’ve read all your answers before posting my previous comment and I checked your repository as well and I uphold my opinion.