One of my applications includes user-generated posts and functions in a similar way to Instagram. When a user opens the app they see a feed of posts sorted by date. This works when there just one small demographic using the app, but as the user base becomes more diverse, not everyone is interested in the same posts. This is why apps like TikTok and Instagram have algorithms to decide which posts to show to a user. Where do I even start with this? I understand that there need to be tags on each post for what they are about (this is where I think I can use machine learning) and then each users information needs to include their interests (I’m not sure what can be used to change this as they like or dislike posts). Is there a simple pre-built way of doing this or any examples? It seems fo be a pretty big secret that mostly big tech companies understand and use.
You could use Google's "cloud vision api(For Images): https://cloud.google.com/vision" and "Video Intelligent Api(For videos): https://cloud.google.com/video-intelligence/docs".
Video Intelligence Api could handle images too from byte stream.
Build a firebase function that analyse posted media with these api.
Build the rest of the logic from here. Find a way to detect their interest from post, save their interests.
are there any detailed information on Google Translate and the GDPR?
In my opinion, translating personal data with the google translate widget is an big issue here, especially if you run an online-store and the user translates pages while checking out (i.e.: last checkout step, where all the personal data including cart-positions, billing-information and user contact-information are preset).
There is a way the exclude parts of the website from being translated (adding "notranslate" class attribute), but i assume the data itself is send to google translate servers anyway?
Looking forward to an answer.
Best regards,
Andrea
Collect some statistics about your visitors and implement localization for major languages they use.
This will probably prevent users from using google translate for your checkout process.
Easily seen in the JSON result from this:
https://www.linkedin.com/countserv/count/share?url=https://www.linkedin.com
Which currently returns:
IN.Tags.Share.handleCount(
{
"count":0,
"fCnt":"0",
"fCntPlusOne":"1",
"url":"https:\/\/www.linkedin.com"
});
Apparently it affects most of the LinkedIn Share buttons/counters out there on the web, including WordPress and other blogs. This has been "broken" since late last week (13 Jan 2018).
I opened a ticket with LinkedIn support. Response was to post here, as this is where the LinkedIn Developers support resides. Hoping for a response that says "Oops, we'll fix this." Or, if deliberately crippled, an announcement that says so. (Twitter made a similar move a few years ago. It was unpopular among developers, but we've moved on.)
Further to Chris Hemedingers response, this feature has now been entirely deprecated
Deprecating the inShare Counter
As you can see they have deprecated this saying:
The share count on its own doesn’t fully reflect the impact that a piece of content delivers, and we encourage publishers and other content creators to leverage the inShare plugin as a way to drive conversation and engage with members on LinkedIn.
They then link to the Documentation saying:
Share on LinkedIn plugin will no longer return share count.
This is massively inconvenient for my company as we have just finished construction of a suite of tools powered by this.
The share count service is back in operation, working as it was before. The outage was deliberate (apparently) but temporary.
As far as I know this is an undocumented API, but it's integral to the "LinkedIn Share" buttons that are used in countless websites/blogs around the world. As such, LinkedIn has no contract/obligation to keep that service running...so consumers of the service in non-LinkedIn components should beware.
Thank you for the update! That was quite frustrating to track this down. I had to research the code, request from the API itself with multiple URLs, submitted a ticket to LinkedIn... and ultimately found myself here and read this. Just a recommendation, I think it would be better to return some kind of error code than a 0. Many people actually display the count on their sites.
Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
Questions asking us to recommend or find a tool, library or favorite off-site resource are off-topic for Stack Overflow as they tend to attract opinionated answers and spam. Instead, describe the problem and what has been done so far to solve it.
Closed 8 years ago.
Improve this question
Background
I'm a one man shop (a micro-ISV). A week after putting my product online I get a mail from one of my customers about a bug. It was an obvious fix and I fixed it in 5 minutes but I realize that the reason why the bug was reported so late is because the only contact I have with my users is through mail.
I feel I need something more but I have difficult time finding the right solution.
I was checking out some solutions, but I would like some feedback from the community
Question
What do you use for a micro-ISV (both online and built into software) when you want to give good quality service and support to your clients?
Have an issue-tracking system that your customers can use through a web page. (You do have a web page, right?) Alternately, if your software is interactive, have a menu entry "Submit Bug Report" which will email you what the user says, and perhaps other useful things (users very frequently omit things like software versions, OS versions, that sort of thing). Or both.
Also, your customers are likely to feel happier if they have a standard way to report problems.
If you want to go beyond the "email us" link, you might consider putting up a bulletin-board or even wiki-style forum on your site for your clients to use. Make your own list of Frequently Asked Questions the first post. I'd recommend using an off-the-shelf package, instead of rolling your own. A pre-existing solution should include the spam-filtering and moderation tools that you'll need.
Another idea would be to start a company blog, and invite users to leave feedback.
You fixed it in five minutes? Sounds like you're already giving good quality service / support. But if you really want a tool, I would check out if Unfuddle.com has a public bug report feature. I love that site.
This is a subject I've thought a lot about (since I'm contemplating doing just what you're doing), and there's considerable precedent for how you could proceed.
Set up a feedback page on your website
Set up a dedicated email account for your website
Set up automated opt-in bug reporting and crash reporting for your software
Set up a twitter account; and conduct twitter searches for your software name
Set up a Google Alert to track when a website or user references your product, and respond to them.
Set up a Uservoice account for your software/website (it's free for a 'small' company).
For a start, you can ensure your website is clear, and has useful sections like FAQs and How-Tos.
Make sure your customers can get in touch with you easily, and that you respond to them in a reasonable amount of time.
If you out and don't have a Blackberry enabled phone you could have your software send you an SMS of the fault.
A well designed website with a forum for news, updates, user discussions is probably a good start. It's worth paying someone to do this for you if you want to spend more time designing and coding good software. The more information you can put out there, the less time you'll spend dealing with customer issues.
In addition to giving your users more options on how to report a problem, your site should also be logging a fair amount of information. Such as, who, when, and what they did.
Further, ANY failure should be logged and automatically be reported back to you. Most clients simply won't say there is a problem and will just move on.
Just basic logging will also give you usability information. What pages do they use the most, which ones are used least, what is different about them. Are there features no one cares about?
Finally, engage your customers by asking them what they would like to see. Quite often their vision is different from yours.
I use ontime as a customer portal and help desk / bug tracking tool. It's free for a one person license. Which is great for me since I'm a one man shop as well. I'm the only full-time employee and have one to two part-time 1099 contractors here and there as work comes and goes.
There are also lots of open source out there. However, I've found the ontime to be dead simple, free for a 1 user license and cheap for 5 user license.
Split your time between development and customer support. If you focus too much on support, new functionality will suffer, and if you focus on development, customers will suffer.
So find a balance and plan portion of your time for development and another part to support.
Also keep in mind that solving the bug is just the first step.
You need to test (preferablyseveral configurations)
create a new installation
possible update manual and help files (and don't forget the translations if it's multi lingual).
Add a new version number (every deliverable must be identifyable).
Update website...
So it often takes several days to ship a single bugfix.
Besides, most customers are happy with a few updates per year. And ocasionally an urgent hotfix if the customer is in serious need of a bugfix.
I have a few systems. My main system is through a fogbugz account with buttons built in to my application that create emails for users so that they can then submit comments / bug reports etc. I also run a wiki as the documentation for my application, although I am the main contributor to the wiki and it does take a lot of effort to keep up to date. Again, there is a menu item in my application that takes users directly to the wiki. I have a built in crash reporter using an open source framework, which again submits emails to fogbugz. Finally I do online video and text based tutorials on my applications website, although I'd like to integrate them more into the application.
One (free) product that I know uses Yahoo Groups (and also a Google Group).
It acts as a mailing list: so if you report a bug, that's seen by other users as well as by the group's owner/moderator (i.e. you).
It also acts as a weblog/archive: so users can search it for known issues/answers before they submit a new message.
Have you tried Casengo? Its a free solution (for 1st agent) for handling email, chat and social media . It might be of interest to you. url: http://www.casengo.com
I am using Casengo for several weeks and is very easy to use.
Jeremy
I am working on a project that requires reliable access to historic feed entries which are not necessarily available in the current feed of the website. I have found several ways to access such data, but none of them give me all the characteristics I need.
Look at this as a brainstorm. I will tell you how much I have found and you can contribute if you have any other ideas.
Google AJAX Feed API - will limit you to 250 items
Unofficial Google Reader API - Perfect but unofficial and therefore unreliable (and perhaps quasi-illegal?). Also, the authentication seems to be tricky.
Spinn3r - Costs a lot of money
Spidering the internet archive at the site of the feed - Lots of complexity, spotty coverage, only useful as a last resort
Yahoo! Feed API or Yahoo! Search BOSS - The first looks more like an aggregator, meaning I'd need a different registration for each feed and the second should give more access to Yahoo's data but I can find no mention of feeds.
(thanks to Lou Franco) Bloglines Sync API - Besides the problem of needing an account and being designed more as an aggregator, it does not have a way to add feeds to the account. So no retrieval of arbitrary feeds. You need to manually add them through the reader first.
Other search engines/blog search/whatever?
This is a really irritating problem as we are talking about semantic information that was once out there, is still (usually) valid, yet is difficult to access reliably, freely and without limits. Anybody know any alternative sources for feed entry goodness?
Bloglines has an API to sync accounts
http://www.bloglines.com/services/api/sync
You have to make an account, subscribe to the feed you want to download, but then then you can download based on Date, which can be way in the past. Not sure of the terms.
The best answer I've found so far, is this: Google reader's unofficial API turns out to have a public access point for their feeds, which means there is no authentication needed. Use is as follows:
http://www.google.com/reader/public/atom/feed/{your feed uri here}?n=1000
replace the text in the squigglies (including the squigglies themselves) with the feed URI you're interested in. More information about the precise arguments can be found here:
http://blog.martindoms.com/2009/10/16/using-the-google-reader-api-part-2/
but remember to use the /public/ url if you don't want to mess with authentication