Client- side (browser) 3D Viewer - browser

I am looking for a browser based viewer for 3D objects. The model(s) are going to be fairly large (millions of faces).
Additional requirements:
GIS capabilities
ability to stream the model
smooth and fluid view
Many thanks :)

SpiderGL has a demo with streaming LoD terrain:
http://spidergl.org/example.php?id=8
This demo shows how to stream a large multiresolution terrain from an USGS dataset managing an internal cache and a request queue.
SpiderGL was created by the Institute of Information Science and Technologies (ISTI), an institute of the Italian National Research Council (CNR).
BTW, the development was some time inactive - current version is now just 0.2.0, but the demo is quite impressive and not to underestimate:

You can use three.js or jsc3d. I've tested with large OBJ and PLY models.

Related

How high will the resolution be for OpenMapTiles Aerial/Satellite Imagery?

So this is more of an inquiry,
I visited https://openmaptiles.com/satellite/ and it is my understanding that "Highres aerial imagery" is still to be added. So, my first question will this imagery comparable to the imagery we see in Google Maps and Bing Maps? I attached some of OpenMapTiles lower resolution imagery for comparison. Also, is there a scheduled release date for coverage of the US? Maybe I'm getting ahead of things here but will the imagery be available for download at a one time fee and will it be as affordable as the OpenStreetMap tiles?
I guess this is related to the cost to acquire high resolution data. At the moment, free data sources such as Sentinel-2, Landsat-8, and CBERS4 are 5m to 30m resolution. Higher resolution data like what one would see in Google Maps and Bing Map are very expensive as they require commercial satellite images or aereal photographs. The gov't and local municipalities often provide high-res geospatial data for free on their websites (like open data), if you need data for only a few urban area.

What are the pros and cons of generating heatmap on back-end/front-end side?

I have a network of sensors that measure air pollution. I want to make a heatmap of this pollution and show it to users on my website.
I'm thinking about the best way to do that. Should the heatmap be generated on back-end or front-end side? What are the pros and cons of both solutions? Any details - like in what format should I send the heatmap to the front-end (maybe using SVG?) would be very appreciated.
I can find these pros for back-end generated heatmap:
Easy integration with other services - like if I wanted for example to send map of air pollution to Slack users I can just download map tiles and put generated heatmap on it.
Ease of making changes - eg. if I want to change a way the heatmap is generated (like colors) I don't have to change it in all other places like web app, Android app, iOS app...
Cons:
Probably more data to send to client.
Additionally, can you name any real-life examples of back-end/front-end generated things that are put on top of the map?
It is a bit hard to answer what the best solution is. It all depends on the data.
if you have say, moderate amount of data points, about 50-100, and prefer having a crisp rendering on a high resolution display, then SVG is the way to go.
if you have hundreds or thousands of data points, rendering those with SVG can be a performance killer on the client side. In this case, render everything to a bitmap on the server side and send it to the client.
Examples:
Google maps (or any other map service) renders maps as bitmap tiles on the server side. The classical Google maps in a browser (e.g. ios Safari showing google maps) will do nothing more. Even street names are rendered to the bitmap. (It worth of noting that on more capable platforms, such as desktop Chrome, things are done more tricky.)
JS side charting solutions will render things on the client side. Looks grand for datasets up to some hundred points. Fails sadly on real life data (for example, working in the stock market industry, examples like "show stock prices of Apple" can make these solutions on their knees. (All price changes of Apple shares from 1990 - including tick level data (so thousands of changes per day) is a different league.
Just for the records, in a recent project after months of development work, we were unable to make a client side canvas based rendering working reasonably well on all platforms, and opted to server-side rendering and solved the problem with 4 days of work.

What language/tool can I use to generate a dynamic 3d model on a website

I am developing some sort of browser (war) game, in which player can have a village with certain different buildings (watchtowers, walls etc..) As an overview I want to generate a small 3d model of the village with little interactivity (maybe they can click on a building in the village and it will display info about this building). It should be dynamic (they can upgrade these buildings, and I would want them to change after an upgrade) and depend on values in the database. What tool or programming language is best to use for this? Thank you!
You could try something like Three.js or BabylonJS.

Where can I find a corpus of search engine queries?

I'm interested in training a question-answering system on top of user-generated search queries but so far it looks like such data is not made available. Are there some research centers or industry labs that have compiled corpora of search-engine queries?
There are a couple of datasets like this:
Yahoo Weboscope:-
http://webscope.sandbox.yahoo.com/catalog.php?datatype=l
Yandex Datasets:-
https://www.kaggle.com/c/yandex-personalized-web-search-challenge/data
A part of Kaggle problem. You can sign up and download.
There are also AOL Query Logs and MSN Query Logs which had been publicised as part of shared tasks in past 10 years. I'm not sure if they are still public. However you can explore a bit.
Weboscope/Kaggle data sets have some specific restrictions. I would suggest the TREC data sets, such as this dataset from 2009

Log transformation to get normal distribution. Applicable for a Lab study?

I did a lab incubation study on nitrous oxide and carbon dioxide emission from different soils. the data are highly skewed. I log transformed the data to run anova. But one of my coworker is really doubtful about that as we haven't found any paper on soil incubation study where researchers used log transformed data. What should I do? Should I log transform my data to get a normal distribution no matter what others do? Our concern is about the reviewers' questions when we submit the article to a journal.

Resources