I'm looking to use Chameleon to send a microsurvey to our users.
Using Segment data as the top of the users funnel, is there an easy way to limit the survey audience to 1% of users?
Related
We have a website with a traditional database (mongodb), which is the user voting system.
Each user can create many votes and we want to migrate that votes to be stored in the blockchain in a secure way, so they can't be changed or deleted once created. There are 300 000 votes in our db atm.
Can we use NEAR to store such an amount of data and how it can be implemented?
And what would be the price for storage?
Yes you can use NEAR to support your voting application.
We actually have an an example of that here
https://github.com/near-examples/voting-app
On top of that we have some useful resources here that discuss storage staking
https://docs.near.org/docs/concepts/storage-staking
That last link on storage will give you an overview of what storage staking is on Near but for an accurate up to date cost of how much storage would cost you per byte in yoctoNEAR (1NEAR=1*10^24 in yoctoNEAR) can query our RPC using this guide
https://docs.near.org/docs/api/rpc/protocol#protocol-config
As well as alternative storage solutions you can use in conjunction with NEAR
https://docs.near.org/docs/concepts/storage-solutions
I have run across some issues while developing MS projects for learning purposes.
So the idea currently is that I have 2 Microservices for now:
User MS (UMS)
Statistic MS (SMS)
But on my UI I have a users page, where are all users displayed but they should be sorted by "Statistic MS" data.
Currently, I have 4000 fake users, and sending new API call for each user is insane. I have a pagination with 30 users per page, even though I'm using caching.
So it means now 30 requests are sent to get the statistics for each user this is a working solution, but very slow, and also it does not sort users based on statistics.
Because I'm getting users sorted by DESC, and just sending API call to statistics service for gathering each user statistics.
But what I need to have:
When someone opens /users page they are automatically sorted from "Statistics Service" so for example users with the best statistics would be at the top.
F.Y.I. The statistics are calculated based on user activity, that's why it's a new MS. But /users page should have a filter for this, so everyone could be able to filter through all the users with the best statistics, etc.
This approach can violate the high availability and is inefficient. For example, if statistics MS is down your user MS will be down too. I think the best approach here can be a self-contained denormalized database. To be more clear, you can add a field in the user's table which determines the value of the statistic for each user.
However, this may add some complexities to your code as you have to keep both microservices (user, statistics) synched together. In this case, you can use a message broker (such as RabbitMQ, Kafka, etc) between the microservices, therefore for each statistic changes in statistics MS, it's going to publish an event on a specific channel which is listened to by the user MS to update its database.
Now you don't need to call the statistics MS for every single request and you can order the users simply by a SQL query. Also the failure of statistics MS will not impact the user MS functionality.
Is it possible to fetch the Google star rating for any business using the Google Places API?
I have a comparison website and want to display the Google star ratings for each business on my site.
Many thanks
Yes. The responses from the Place Search and Place Details APIs include a rating field.
However, two important warnings:
These APIs are both billed, and are quite expensive ($17 and $32 per 1000 requests, respectively). Making a Place Details request for each business displayed in a comparison will probably be economically infeasible.
The Places API policies place a number of requirements on your use of Google's data. In particular, you cannot cache most data returned by the API (including ratings), and you cannot use the data alongside a non-Google map.
We are using Parse to send out Push notifications using the Parse REST API. We compute the audience of the push notification based on dynamic user data such as user's current location. And in our Production system, we observe that there are times when this user base can be quite high given the time of the day. During such times, we have seen the ParseException:
org.parse4j.ParseException: Neither where clause nor data may exceed 64KB
This is because the where clause has a large number of "installation ids" or "device tokens" specified as we find a large number of users in a given location.
I understand that Channels / Parse Audience is a way to deal with larger sets of users. But this requires me to store the dynamic data like a user's current location in the Parse database as part of the Installation metadata.
My questions are:
Is it the right way to implement it if we decide to store user location in Parse, which will also mean that we would need to update this Installation object per user very frequently.
Is it advisable to just send push notifications via Parse in chunks, that is first to a set of 2000 users, then the next 1000 etc.
Is there any other way to handle such a case?
From my understanding, a bounced session is a session with one page view only. However, today on Google Analytics I chose the following two segments:
Bounced Sessions
Custom Segment; Filter used Pageviews = 1
The number of sessions for the custom segment was nearly 20% higher than the number of sessions for the Bounced Sessions segment.
I'm confused as to why both segments didn't give me the same number of sessions. Do I have the concept of bounce rate messed up? If so could you help me understand better what Google Analytics counts as a bounce?
Bounce session - is session with only one GET query to the one to one pixel file on google-analytics.com. Сonsequences are follows.