how to interpret "Data" tab of BRouter web client

[Andrew Heard]

Is there any doco on how to interpret the "Data" tab of BRouter web client? I am trying to understand more deeply why one profile choose track A, and other profile chooses track B.

It would be nice if the text in the Data tab could be exported in CSV format for importing into a spreadsheet for further analysis. PS - unnecessary - I can select the text, then copy/paste into spreadsheet. But I leave the question here as maybe useful tip for others.

For example, is cost per row = dist/1000 x $/km + elev$ + turn$ + node$ + initial$
And therefore total track cost = sum of (cost per row).

Thanks for any help.

[Poutnik]

Dne 26/05/2016 v 02:23 Andrew Heard napsal(a):

> Is there any doco on how to interpret the "Data" tab of BRouter web
> client? I am trying to understand more deeply why one profile choose
> track A, and other profile chooses track B.

I guess you have described it quite good.

The most important is comparing
cost per km and way tags on the right for the given way segment.

And switching profiles and watching how the cost changes

>
> It would be nice if the text in the Data tab could be exported in CSV
> format for importing into a spreadsheet for further analysis.

Why not to usen Data CSV link just above the tabs ??


--
Poutnik ( The Wanderer )

My Brouter profiles
https://github.com/poutnikl/Brouter-profiles/wiki

[Poutnik]

Dne 26/05/2016 v 02:23 Andrew Heard napsal(a):
>

> It would be nice if the text in the Data tab could be exported in CSV
> format for importing into a spreadsheet for further analysis.
>

Be aware that Brouter-web is aimed on desktop browsers,
it is mobile and small screen unfriendly. ( but changes are planned )

In many of Android browsers,
some of its feature do not work, or work strangely.

Once I did some testing with 1-2 users
and we have realized that
all features work ain Android browser,
but never all at the same time. :-D

E.g. in one browser works A, B but not C
in another works B, and C, but not A.

[Poutnik]

As you may know, Brouter always try to find "shortest route".
"Shortest", as all priorities and penalties
are recalculated as effective length of a way segment.

3 parameters are important for a way segment evaluation
Costfactor - is the profile evaluation of a way quality for the profile
purpose. costfactor >=10000 mean forbidden road, excluded from
consideration.
Turncost - is additional cost for every turn, it is effective length
addition of sharp right angle turn ( Angle dependent, 0 .. 2*turncost )
Initialcost - is additional one time cost of way segment of given
quality. Zero physical length has Initialcost effective length.

Additionally, OSM nodes ( points of a special meaning )
have Initialcost as well ( e.g. traffic lights can have, or stop sign,
or toll booth etc )

Raw way segment effective length = physical way segment length * Costfactor.
E.g. 10 km of the bestroad is taken as 10 km, while 10 km of a bad road
can be taken as e.g. 30 km.

Then there is added Turncost and InitialCost

For elevation, the cutoff is subtracted from the slope, it is then
multiplied by the segment length, resulting in the effective elevation
difference.
this difference multiplied by elevation cost gives additional length to
the effective length. ( this is more or less valid
for larger elevation differences and way length. For small ones and
local changes, it is complicated by the elevation buffer functionality,
what I will left aside now ).

finally Node initial costs are added.

All is summed along the route , giving final cost, what is taken as
effective route length.
the Brouter choose the one with the lowest cost.

Poutnik

> --
> You received this message because you are subscribed to the Google
> Groups "OSM Android bikerouting" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to osm-android-biker...@googlegroups.com
> <mailto:osm-android-biker...@googlegroups.com>.
> For more options, visit https://groups.google.com/d/optout.

[Andrew Heard]

Thanks for the huge reply.

I had not noticed the "data CSV" link - also thanks. Although I note the downloaded file format is Tab Separated Values (TSV), not Comma Separated Values (CSV).

[Poutnik]

Well, in these days is TSV usually considered
as CSV ( in broader sense ) with delimiter character = tab.

As CSV is this broader sense is considered primarily
not as delimiter definition,
as data format for tables:

1 data being in a plain text form,
2/ rows delited by line OS delimiters
( CRLF for Windows, LF for Unix/Linux, CR for Mac )
3/ items delimited by item delimiters.
( comma / tab / semicolon )

On 05/27/2016 09:51 AM, Andrew Heard wrote:
> Thanks for the huge reply.
>
> I had not noticed the "data CSV" link - also thanks. Although I note the

> downloaded file format is _Tab_ Separated Values (TSV), not _Comma_
> Separated Values (CSV).
>