Lessons Learned from Suggesting 72 Road Edits

[MortenCopenhagen]

This is part 3 in a mini-series with tips on Editing Roads. Find the series index in reply # 1 below.

While Part 1 and Part 2 described methods for keeping track of individual road edits, this post presents the results of making 72 road edits between February 12th and 15th 2024.

Hopefully, these results will reveal what types of road edits are more likely to be approved. Or at least form the basis for deciding if adding roads is worth my time. Should the approval rate of these road edits be very low, then this can be explained by either:

1. My poor ability to locate, make, and predict the outcome of road edits

or

1. The AI system and the operators approving road edits are at fault

or a combination of the two.

72 road edits is a very small sample, but the time needed to document, and track the outcomes daily took a lot of time. Also, this sample is limited to the kinds of road edits I was able to find. Where I live, most municipalities upload their detailed geo data regularly to Google Maps, so finding missing or incorrect roads in Denmark is close to impossible.

RESULTS

Let’s take a look at the 72 road edits made over 3 days in February of 2014.

Number of edits per day
Date	COUNT
Feb 12	36
Feb 13	17
Feb 15	19
Grand Total	72

As mentioned above finding roads in need of editing in Denmark is not easy, so I decided to also search in 6 other countries.

Number of edits per country
Country	COUNT
Denmark	6
Latvia	15
Portugal	2
South Africa	13
Sweden	31
Uruguay	3
Venezuela	2
Grand Total	72

We can add new roads, delete them, rename them, adjust the alignment (fix the shape), etc. In this sample, I mostly added missing roads as seen here:

Type of road edits
Edit type	COUNT
Add	62
Remove	6
Shape	4
Grand Total	72

As most roads are already on Google Maps 2 out of 3 edits were missing parking lots or missing roads inside parking lots.

Parking	67%
Walking/Biking paths	21%
Local roads	18%

I wanted to know if gravel roads and parking lots are less likely to get approved, so I also registered if the road edits were on a paved surface.

Paved surface
Paved	COUNT
Yes	52
No	20
Grand Total	72

A road edit with only one line was classified as “low” complexity. While a lot of roads and intersections in one edit were classified as “high” complexity. Road edits with up to 5 individual roads were “medium” complexity.

Complexity
Complexity	COUNT
High	15
Low	22
Medium	35

As seen above I mainly made road edits of medium complexity but also quite a few high and low complexity edits.

The most important outcome is whether the road edits get approved or get rejected. At the time of writing this report, all road edits from February 12th, 13th, and 15th have been resolved as approved or rejected.

Overall approved rate
Status	COUNT
APPROVED	17
REJECTED	55
Grand Total	72

The main result is that only 24% out of 72 road edits got approved. So 3 out of every 4 road edits were a waste of time!

This is a very disappointing result. It is not surprising, as I suspected this could be the case.

From earlier road edits I expected this test could take several weeks as Road edits could stay pending for many weeks. But to my surprise, the delay time has been minimized by a lot.

How quickly are road edits processed?

All edits suggested on Feb 12th were decided (rejected or approved) within a week. (7 days).
All edits suggested on Feb 13th were decided (rejected or approved) within a week. (6 days).
All edits suggested on Feb 15th were decided (rejected or approved) within a week (7 days).

This is really good news.

Before looking into the individual edits, let us check if there are differences in the approval rate depending on the country, type of edit, type of road, and road surface.

Approval rates over countries
	Status
Country	APPROVED	REJECTED
Denmark		100%
Latvia	27%	73%
Portugal		100%
South Africa	23%	77%
Sweden	29%	71%
Uruguay	33%	67%
Venezuela		100%
Grand Total	24%	76%

As you can see none of my edits made in Denmark, Portugal, and Venezuela got approved. The approval rates in Latvia, South Africa, Sweden, and Uruguay are very similar: Only about one in three edits got approved.

This confirms I should stop trying to add roads in my home country!

Approval rates over types of edits
	Status
Edit type	APPROVED	REJECTED
Add	27%	73%
Remove		100%
Shape		100%

Trying to remove roads or realigning roads is not my thing. 100% of these edits got rejected. So should I want to improve my approval rate, I should probably stick to adding missing roads.

Approval rates over edit types
	Status
Rd type	APPROVED	REJECTED
Parking	32%	68%
Path	8%	92%
Local rd	17%	83%

The success rate for adding parking is higher than editing walking/biking paths and Local roads.

Finally, let’s see if the complexity affects the approval rate of road edits.

Approval rates over complexity
	Status
Complexity	APPROVED	REJECTED
High	7%	93%
Low	36%	64%
Medium	23%	77%

It sure does. The lower the complexity is associated with the highest approval rate. And it is very unlikely for very complex edits to get approved. This is an important finding.

Finally, I would like to report on the speed at which the road edits get approved. The average delay is 31 hours from the initial acknowledgment email until we receive an email confirmation of approval. As the approval emails are only shipped once a day (between 2 and 3 pm Danish time) the average delay is likely shorter. I did find approved edits during the manual check before the email arrived. Unfortunately, I did not record the delays before road edits were rejected. But all pending edits were as mentioned above resolved within one week.

31:11:28	avg
5:17:00	min
153:06:00	max

If I were to redo this test I would also have recorded data on the visibility of the roads in satellite mode. Are all parts of the road clearly visible or are there trees, bridges, buildings, shadows, etc blocking the view of the road? A relevant road edit to discuss this is “Feb12 11”.

DISCUSSION

So now the biggest question is whether the low approval rate of 1 in 4 (24%) is sufficient to justify our time spent on editing roads. How we value our time and how low an approval rate we are willing to accept is a personal choice. However, we can discuss the following underlying questions:

1. Is the low approval rate of 24% caused me being a newbie and inexperienced at spotting and making road edits, or is the approval system wrongfully rejecting perfectly good road edits?

and

1. What can we do to increase the approval rate of road edits?

Personally, wasting my time in 3 out of 4 road edits is not acceptable. I want a higher success rate to be motivated to keep doing road edits as a volunteer.

To answer question A I would like to invite follow road editors to share their approval rates in the comments below. This could prove to be an assurance that it is possible to do better.

We know from the filters approving photos and reviews that filters can be overly sensitive leading to wrongfully removed photos and reviews. I speculate that approving road edits is mostly done by human operators and not an AI system. I conclude this based on the fact that road edits in my experience are never approved immediately. But even human operators can be instructed to not accept edits if in any doubt.

As confirmed in this test, I have had huge issues with getting road edits approved here in Denmark. I believe a lot fewer than 24% of my prior road edits in Denmark got approved. So on that note getting 1 in 4 road edits approved is an improvement for me. This might be because most municipalities are geo data upload partners and such official data sources are trusted over maps users. From this test, I think it is safe to conclude that the approval rate of road edits depends on which country or region you are making the edits in.

To answer question B: What can we do to increase the approval rate? I would like to invite fellow editors to critically review my 72 road edits documented below. And please share concrete suggestions and tips on getting a higher approval rate. I made them as perfect as I could, so I sincerely hope fellow (and more experienced) road editors will evaluate and spot the blunders I have been making.

For this purpose you will find detailed information for each edit I made. There is a screenshot followed by the reference to the road edit like “Feb 22 1” which indicates that this is the first road edit made on February 22nd. Please use these references when making comments and suggestions. The outcome APPROVED or REJECTED is also mentioned for each road edit. Further, the type of edit and type of road is indicated. And lastly, there is a Maps link so you can check if you think a road is missing, or needs adjustment or removal.

I’m sharing all this information directly here in this post rather than asking you to open and navigate the spreadsheet. Should you want to see the spreadsheet I will be happy to share. It might be useful as inspiration should you plan to systematically document the outcome of some of your road edits.

I’m hoping for a lively discussion. Don’t hold back.

Here are two examples of what concrete feedback could look like:

Re Feb12 35

Why did you submit so many roads in one edit? You should split this parking lot up into several road edits. First suggest the outer perimeter added. Then after this is approved you could add the “internal” roads one by one. Instead, you could start with this:

Re Feb12 19

Why would you want to remove these two half-circles? They could help drivers know where to make a U-turn?

At the end of this long post, you will find 5 road edits that I went back and re-submitted. All to no avail.

I hope this post and the following discussion will be helpful to other aspiring road editors.

All the best

Morten

Feb12 1 REJECTED Add Parking

https://maps.app.goo.gl/oVCHA7eQzviPUUub7

Feb12 2 REJECTED Shape Path

https://maps.app.goo.gl/WD5QCKQ2vtkpyzHp9

Feb12 3 REJECTED Add Parking

https://maps.app.goo.gl/hWvqhdq28fu7iPuN7

Feb12 4 REJECTED Add Parking & Path

https://maps.app.goo.gl/b7aeccAxn9k7XkXu6

Feb12 5 APPROVED Add Parking

https://maps.app.goo.gl/M6AcmXdpuZ8LNxJZ6

Feb12 6 APPROVED Add Parking

https://maps.app.goo.gl/t8XNNC7UuZiMXrGRA

Feb12 7 REJECTED Add Parking

https://maps.app.goo.gl/cp2YvJpmYjhGZaX69

Feb12 8 APPROVED Add Parking

https://maps.app.goo.gl/QtaYE2sC8Qege5np6

Feb12 9 REJECTED Add Parking

https://maps.app.goo.gl/9Hc7v2vi9k7ABWiY6

Feb12 10 REJECTED Add Path

https://maps.app.goo.gl/B9DDcwyhaTENLv1d7

Feb12 11 REJECTED Add Path

https://maps.app.goo.gl/t27rcoXWzNabzETk7

Feb12 12 REJECTED Add Parking

https://maps.app.goo.gl/sjSu2kzy3MeR1GAB7

Feb12 13 REJECTED Add Parking & Path

https://maps.app.goo.gl/gZD15wyhNocyUjA19

Feb12 14 REJECTED Add Path

https://maps.app.goo.gl/6uzTkGw7eEyZDpQU6

Feb12 15 REJECTED Shape Local rd

https://maps.app.goo.gl/kEHEVRQnbyMiHgmz9

Feb12 16 REJECTED Add Local rd

https://maps.app.goo.gl/MbtGvscs8oPFYcbM7

Feb12 17 APPROVED Add Local rd

https://maps.app.goo.gl/Gqr5PuEUgRTvhh5S9

Feb12 18 REJECTED Remove Local rd & Path

https://maps.app.goo.gl/d2kaLjRNCKiV7PvH7

Feb12 19 REJECTED Remove Local rd

https://maps.app.goo.gl/reuooa7FpDS7NQad8

Feb12 20 APPROVED Add Local rd

https://maps.app.goo.gl/PFihXj3HWRcEAT34A

Feb12 21 REJECTED Remove Path

https://maps.app.goo.gl/Rbn9cZAex1k5c5846

Feb12 22 APPROVED Add Parking

https://maps.app.goo.gl/n5HQFmppbGWVKFdm8

Feb12 23 REJECTED Add Parking

https://maps.app.goo.gl/RtsUyhj39L3WEiqYA

Feb12 24 REJECTED Add Parking

https://maps.app.goo.gl/ExzVRB4MmDP54HMA8

Feb12 25 APPROVED Add Parking

https://maps.app.goo.gl/yHox8dGn4rCcRvTY7

Feb12 26 REJECTED Add Local rd

https://maps.app.goo.gl/gsmZ6PFBx9Rc6K8x7

Feb12 27 REJECTED Add Parking

https://maps.app.goo.gl/S4wRGZgS8G6S2muX9

Feb12 28 REJECTED Add Parking

https://maps.app.goo.gl/jMdVpU5N51oQvNid7

Feb12 29 REJECTED Add Parking

https://maps.app.goo.gl/p4S4rp5Enr7ipX3M9

Feb12 30 REJECTED Add Parking

https://maps.app.goo.gl/Lv6MmwLt9xDfdp6X9

Feb12 31 REJECTED Remove Local rd

https://maps.app.goo.gl/zqhYgNxm5cASQQKs5

Feb12 32 REJECTED Add Parking

https://maps.app.goo.gl/nZaC1M2oNcZz5a7f7

Feb12 33 APPROVED Add Parking

https://maps.app.goo.gl/sdyutzWxZbyd7SKq8

Feb12 34 REJECTED Add Path

https://maps.app.goo.gl/RUo6Pg1UgMHLFXhH9

Feb12 35 REJECTED Add Parking

https://maps.app.goo.gl/n5RkViAxibtgWqKj6

Feb12 36 REJECTED Add Parking

https://maps.app.goo.gl/3VKvJystowLhGaam9

Feb13 1 REJECTED Add Parking

https://maps.app.goo.gl/R9dryLwV3pdVaCcK6

Feb13 2 REJECTED Add Parking

https://maps.app.goo.gl/euJDV7aZ6u7s93Ax8

Feb13 3 REJECTED Add Parking

https://maps.app.goo.gl/w6HvaGf79HdHYFUk9

Feb13 4 REJECTED Add Parking

https://maps.app.goo.gl/x9bdXcqWbmrFKmuCA

Feb13 5 REJECTED Add Parking

https://maps.app.goo.gl/oevE5sqJq1ZpKxwL7

Feb13 6 REJECTED Shape Parking

https://maps.app.goo.gl/rRTUyKyWhUZkMAer7

Feb13 7 REJECTED Add Parking

https://maps.app.goo.gl/tMoNmkBhmtJE9ZX46

Feb13 8 APPROVED Add Parking

https://maps.app.goo.gl/cTRBSf1FimQivRBH7

Feb13 9 REJECTED Add Path

https://maps.app.goo.gl/SWoqh2th9SpvSWei6

Feb13 10 APPROVED Add Parking

https://maps.app.goo.gl/PYEWjZzuTaKGgquT9

Feb13 11 REJECTED Add Parking

https://maps.app.goo.gl/jHcxps6rJtfhd8ri9

Feb13 12 REJECTED Add Path

https://maps.app.goo.gl/mW9wAr7eSbhGhx9AA

Feb13 13 REJECTED Add Local rd

https://maps.app.goo.gl/GmD5X8BBzHbnCBP69

Feb13 14 REJECTED Add Path

https://maps.app.goo.gl/EsoBRWMdCD7TCovX9

Feb13 15 REJECTED Add Parking

https://maps.app.goo.gl/NpsstraekwgAqkAg7

Feb13 16 REJECTED Add Parking

https://maps.app.goo.gl/oVcDHVUM1VTkyPg27

Feb13 17 REJECTED Add Path

https://maps.app.goo.gl/wvrEejdRWQyvvnyU8

Feb15 1 REJECTED Add Parking

https://maps.app.goo.gl/VCtpqpCokRxkKPRm9

Feb15 2 APPROVED Add Parking

https://maps.app.goo.gl/Mxv7y2DYtdeYi6BA9

Feb15 3 APPROVED Add Parking

https://maps.app.goo.gl/VCtpqpCokRxkKPRm9

Feb15 4 APPROVED Add Parking

https://maps.app.goo.gl/jnnBWS34dfgSyAmL9

Feb15 5 REJECTED Add Parking

https://maps.app.goo.gl/YCnn35Sirm3xtGMK9

Feb15 6 REJECTED Add Parking

https://maps.app.goo.gl/uByG9gtLutoeVJkP6

Feb15 7 APPROVED Add Path

https://maps.app.goo.gl/mnB6SKdvb1EnPSpf7

Feb15 8 REJECTED Add Parking

https://maps.app.goo.gl/XZ6MZiWWGgTT48QbA

Feb15 9 APPROVED Add Parking

https://maps.app.goo.gl/G7XBuCPcH4duUBbU7

Feb15 10 REJECTED Add Local rd

https://maps.app.goo.gl/awAAnDp2nAvtTt1S6

Feb15 11 APPROVED Add Parking

https://maps.app.goo.gl/vzJJQnTzaZjn7iFa7

Feb15 12 REJECTED Add Parking

https://maps.app.goo.gl/Y5BDykEKDGTyvs8h6

Feb15 13 REJECTED Add Parking

https://maps.app.goo.gl/XNt2QNDNKF9zicv6A

Feb15 14 APPROVED Add Parking

https://maps.app.goo.gl/fgWM46fCeJimBu2r6

Feb15 15 PENDING Remove Local rd

https://maps.app.goo.gl/MrG2twWf48MPDDeT6

Feb15 16 REJECTED Add Parking & Local rd

https://maps.app.goo.gl/MrG2twWf48MPDDeT6

Feb15 17 REJECTED Remove Path

https://maps.app.goo.gl/PHGeWK4WkmNhGcrq9

Feb15 18 REJECTED Add Local rd

https://maps.app.goo.gl/qJ1FmtUSage8p1rPA

Feb15 19 REJECTED Shape Local rd

https://maps.app.goo.gl/gfe1AGDFvF4aX1Kc9

TRY AGAIN

I attempted to redo the following 5 road edits to no avail.

Redo of Feb 12 4 Also rejected

Redo of Feb 12 4 also rejected

Redo of Feb 12 35 also rejected.

Redo of Feb 13 9 Bikepaths on the north edge not included this time. Also rejected.

Redo of Feb 1 12. This time with fewer paths. Also rejected.

Thanks for reading my post all the way to the end.

Cheers

Morten

[MortenCopenhagen]

Tips for Road Editors mini-series Index:

Part 1: Creating Gmail Folders and Filters

Part 2: How to Keep Track of Your Road Edits

Part 3: Lessons Learned from Suggesting 72 Road Edits

Part 4: Manual for Road Editors

Part 5: Conclusions

Cheers

Morten

[MortenCopenhagen]

Conclusions

Please see part 4:

Manual for Road Editors

[MarcoDavoli]

Dear @MortenCopenhagen ,

I went trough “Part 1 and 2” and now I enjoined the reading of Part three.

I’m sure you’re aware that with your statement: "So now the biggest question is whether the low approval rate of 1 in 4 (24%) is sufficient to justify our time spent on editing roads. " you will ignite a big debate among the most ardent “roaders”. You’ve raised some well-known points.

I wanted to respond to you “writing straight away” but tonight, calmly, I will try to give you my perspective.

Thank you for “stimulating” this discussion.

Ciao

Marco

[MortenCopenhagen]

Hi @MarcoDavoli

I feel that I somehow open myself up to a lot of criticism. But I hope I can handle it. The purpose is to enlighten us all and learn together.

I’m looking forward to your comments.

Earlier today I learned that @ErmesT can get 98% of his road edits approved - so I know there is a huge potential for improvement.

Cheers

Morten

[PrasadVR]

Connect Moderator @MortenCopenhagen I read the Part-1 and 2 posts. Thank you for sharing the results of your research work in this Part-3 with tables and screenshots. I appreciate your hard work on this.

I would like tag our friends @JustJake , @Lukas_1 , @Rednewt74

[JustJake]

@MortenCopenhagen My first thought was where you stood in the “trust factor” of adding roads, but I see that you’ve already gotten over 150 approved. It seems that each type of these edits - adds, reshaping, deletions - requires a separate level of “trust” and approval assigned to you. It also seems like the type of road (parking lot, walkway, highway, local road) has a certain level of “trust” assigned to it.

Just like you can probably get a POI edit approved faster than most, this may be a case where starting slow will be necessary to prove to our Googley Watchers that you know what you’re doing.

I will point out that deleting a road is inherently difficult. I refer back to the Q&A on Road Mapper where Katie McBroom explained that roads should be fixed, not deleted. I tend to use the ‘Other’ option when realigning a long road to the satellite image, reassigning a road to another type, or removing a major road violation.

Deletions should be reserved for only small segments of roads that are blatantly incorrect.

I’ll let @ErmesT opine on your approval rates and screenshots. It does stand out to me that you were able to have all of these roads resolved within such a short time. Even after 6,000 roads added, and over 1,000 edited, I still have times that a decision to approve or reject takes weeks to months.

And, to address the “waste of time” part … yes, spending hours finding and drawing parking lots just to get the edit rejected may feel like you’ve wasted time that could have been spent doing other edits or projects. However, I would disagree that finding and connecting homes, villages, and people that are unmapped is never a waste of time. It might be disheartening to see the low level of approval, but each successful connection is bringing a benefit to the area. I have never judged this as a waste of time. Edit or don’t, that’s up to you, but calling it a “waste” is a little offensive to the mission that many of us are on.

Thank you for this series. The discussion needs to be had. We are about to head into another TRAC event and having the tools you’re helping explain will be a great benefit for those involved.

[MortenCopenhagen]

Hi @JustJake

Thanks for your reply.

I’m still hopeful that a better understanding of how the approvals are done will allow a much higher approval rate. Fingers crossed that Mr 98% approval rate aka @ErmesT can share some tips for this that effectively can penetrate the trust score theory. To test the trust score concept I could stop all road realigns and removals. Interesting proposal to use the other field. Do you have positive results from doing so?

All the best and thanks

Morten

[ErmesT]

Thank you for this post, @MortenCopenhagen , that is basically confirming what I wrote exactly 10 months ago: 3000+, and counting. My tips for adding roads in Google Maps . 25 thousand road later (yes, this is the amount of road Google Maps approved on my profile on 10 months) I can confirm the content of the post.

Honestly I don’t believe that there’s a “trust” score in adding roads, @JustJake , and this is for a good reason. The approval process is based on two steps:

• Preliminary verification by an AI. The AI has the power to reject the edit or to put it in a stack for the final approval by a human operator. In the stack the roads are ordered by priority.
• Final approval by a human operator. It may takes from a few hours up to 90 days, depending of the priority of the edit.

The first check made by the AI evaluate the edit. If the priority assigned to the edit is extremely low, even if not necessarily wrong, the edit could be immediately rejected, for the simple reason that it would not be processed in the time frame of 90 days. This part includes modification of shapes only, and complex edits.

• Shapes. While at the beginning of the program the modification of a shape was accepted (few contributions and only human evaluation) when the program was opened to everyone the priority of shapes was lowered a lot, for at least a couple of reasons. Not a real benefit in the usage of Google Maps, because the route remains basically the same, and an opportunity for the gamers to get points by simply touching a point of a road. In the past I wrote a post about correcting shapes, and why it is important to do that.
• Complex edits. Again, it is a matter of Costs Vs Benefits. Not being a fanatic of points, at the beginning of the program I was adding a whole neighbourhood, or small villages, on a single submission. You can read my post about that. But complex edits are difficult to evaluate for the AI, and the approval could take a very long time. So even if it is more “time consuming” for us to build an area in small steps, you can achieve the result in a shorter time than with complex edits.

I am not surprised at all about your rejections, Morten. To increase your approval rate you should think like the AI. Editing a shape only is a big no, while the edit of a shape together with the addition of a road is normally approved quickly, if the new road is connected to the shape, as you can see here below

A complex edits is a big no too. I usually makes an edit of a big area in two or three steps:

1. Submission of the main roads
2. Submission (one by one) of the internal roads. You can add more than a road if the two roads are connected, but avoid to add more than two connect roads in a single submission. On this way you can easily submit up to 5 roads in one single minute.

Priority: the AI prioritize edits in areas where more users can benefit of the new roads (high density) so edits in remote areas are getting low priority. The same principle applies for pathways and bicycles road. I complained already about this choice because I think that Google should prioritise a sustainable mobility, and I will write a post soon about that.

In addition frequently the bicycle roads are very close to the regular one. This is making difficult for the AI to separate the lines.

There is much more to say, but I think my reply is already too long and difficult to read

Just a note about the 98% of approvals: the number doesn’t consider the edit rejected after 90 days, because on that case the rejection is not due to a wrong edit. I am preparing a detailed post about the subject, and how Google can act to avoid it. In fact I submit again all the edits rejected for that reason, and they are usually approved in a very short time.

[JustJake]

@ErmesT Regarding the “Trust” factor, anecdotally, we saw/see many LGs adding their first roads having approvals taking weeks to be acknowledged. Now, as you know, approval/rejection can be rapid-fire. I’m only left to assume that there is more scrutiny initially.

[JustJake]

@MortenCopenhagen Yes, the ‘Other’ option has been very successful for me. Lately, I’ve been using it for areas where it’s clear the road layer did not get properly aligned with an update to the satellite image. I’m not sure if keywords play a part, but I’m always clear about what thing the mapped road is violating, such as homes, properties, physics … It takes some time, but the fixers are very good at redrawing.

[ErmesT]

Well, honestly it is more a lack of strategy, @JustJake , and this it is absolutely normal when you add your first road.

Me too, and I think all of us, when we added our first road we just added a (one) missing road close to our place.

So probably the road had a low priority.

As you know one of the key factors to have your road checked is the density of edits. Usually Google Maps approves edits in an area all together.

But of course a high number of approved roads could be one of the factors for prioritising an edit. I have no information about that, and some of my edits are approved after two or more months from the submission.

@MortenCopenhagen , I am of course checking the places you shared, with some interesting result, but I need a few more days to makes them more consistent. I think you may have a lot of work to do in that area. Let see what Maps will say on the next few days

[MarcoDavoli]

Dear @MortenCopenhagen ,

grazie per il post ed il pensiero che hai voluto condividere. Sono convinto che il tuo post sarà estremamente utile per coloro che desiderano immergersi nel meraviglioso mondo dei “Roaders”, contribuendo alla creazione di nuove strade.

Senza dubbio, la tua “mini serie” affronta tematiche alla ricerca di idee e soluzioni che solo chi è un “Roaders addicted” può davvero comprendere.

Come saprai, ho seguito con interesse le campagne TRAC proposte dal nostro amico @SholaIB , durante le quali ho potuto contribuire con creazione di oltre 16.000 nuove strade e, con molta umiltà, forse un pò di esperienza me la sta facendo.

Partecipare a queste campagne offre l’opportunità di scambiare idee e suggerimenti in tempo reale per migliorare la nostra attività. A tal proposito invito tutti a leggere il nuovo post di Shola di presentazione del prossimo Meet-up.

Proverò a fornire un commento dettagliato sulla tua mini serie sperando di portare alcuni contributi

.

“Roads approved, refused or pending”

La cartella Gmail - è una soluzione eccellente che ho adottato anch’io, insieme a molti altri “roaders”. È vantaggiosa per separare le email e per gestire facilmente la contabilità del nostro lavoro. Personalmente, ho aggiunto un passaggio alla mia contabilità: ho creato una semplice tabella Excel

Questa tabella riporta solo una parte dei dati per motivi di lunghezza, ma mostra un quadro significativo.

Nei mesi di novembre, dicembre e gennaio, il sistema ha approvato ben 7025 nuove strade, tuttavia ne ha scartate oltre 6000, quasi il 50% di quelle inviate.

É inevitabile domandarsi, nonostante le migliaia di “submission”, perché solo così poche venano approvate.

Personalmente, ritengo che questo dato sia difficilmente giustificabile.

Per cercare di migliorare questo triste primato, ho condotto varie prove, chiesto feedback a Google e coinvolto molte persone. Tuttavia, fino ad oggi, non ho ottenuto risposte definitive, solo sensazioni e supposizioni. Alcune delle risposte che ho ricevuto includono:

• La gestione delle modifiche è affidata all’intelligenza artificiale, che a volte può commettere errori.
  • Anche a me sono state rifiutate alcune modifiche / creazioni di nuove strade del tutto legittime. Infatti, al secondo tentativo, sono state approvate.
• La presenza di ombre e alberi può confondere il sistema, portando al rifiuto delle modifiche.
• Le modifiche potrebbero riguardare tratti che attraversano proprietà private.
• In alcuni casi, le modifiche proposte potrebbero già essere state inviate da altri utenti
  • una sfortunata coincidenza!
• È stato notato che il lavoro viene cancellato dopo un certo periodo, anche se sono trascorsi solo tre mesi.
  • Ma perché? Questa è una domanda senza risposta.

Oppure, come dice @JustJake , potrebbe essere dovuto ad un “ level of trust and approval assigned to each of us”. (could be ).

In ogni caso, come tu scrivi, aspettiamo I suggerimenti di Mr. 98%; @ErmesT . Sono certo che avrà preziosi consigli su come migliorarci.

Quando posso dedico del tempo a rivedere, nella sezione “Your contributions → Edits”, quelle modifiche che sono state “not accepted”.

Anche qui il sistema potrebbe migliorare sensibilmente dandoci una mano.

• Mi sono chiesto più volte perché non viene riportato il link della modifica originale?
• Perché non è possibile selezionare tra “Accepted” e “Not Accepted” ?

Per chi è anche attivo su “RoadMapper” sa che qui c’è il link alla modifica apportata. Quindi non è un problema tecnico.

Magari va solo segnalato e questa opzione verrà aggiunta?

Queste sono due schermate di RoadMapper

Tutte queste attività sono svolte con l’obbiettivo di “to become a better Road Editor and achieve a greater approval rate on road edits” (as you wrote)

Alla fine, per ovviare a queste mancanze, quando effettuo delle modifiche tengo nota delle coordinate GPS con qualche annotazioni.

Queste che seguono sono alcuni esempi di lavori svolti il 23 febbraio

e da pochi giorni ho iniziato anche a verificare le date in cui le mie modifiche erano state inviate ed approvate. Con grande sorpresa sono quasi tutte modifiche approvate nel giro di uno o due giorni !

Se mi devo basare sul mi triste primato del 55% di “approval rate”, paragonato al tuo 24%, dovrei smettere immediatamente! Ma resto convinto che la pratica migliori e che, forse un giorno, qualcuno di Google vorrà dare delle spiegazioni e/o aiuti ufficiali.

Cosa fare per aumentare il rateo di approvazioni?

Ho creato due nuovi account per fare delle prove. Qualche mese fa mi ero concentrato su questi due paesi: Libia e Marocco. La maggior parte delle strade non erano asfaltate. Il rateo di approvazione è stato disastroso: 25-30% di modifiche approvate!

Quando mi sono spostato in un’altra area (Uzbekistan 40.99938456056498, 71.08332423562432) dove le strade, per quanto piccole, erano in cemento o asfalto. Il mio rateo di approvazioni è salito di molto. Oltre l’80%.

Posso anche dire che la creazione di una strada nuova che unisce due strade è quasi sempre approvata. Contrariamente, una strada nuova che finisce nel “nulla” spesso non viene approvata.

Una strada molto lunga verrà difficilmente approvata. So che Ermes ha avuto successo con delle strade molto lunghe in Argentina.

Io ho trovato la nuova “Ashgabat-Turkmenabat high-speed highway” Una nuova fantastica superstrada; ben oltre 600Km tutti da registrare…

Vi ho dedicato due giorni. Ora ho deciso di gettare la spugna e provare “Other Option”. Speriamo.

Quando scrivi: “Is the low approval rate of 24% caused me being a newbie and inexperienced at spotting and making road edits, or is the approval system wrongfully rejecting perfectly good road edits?” capisco cosa si provi.

Qualche volta le mie strade sono un pò azzardate ma accetto il rischio (ie: molti alberi, ombre o strade nel deserto) ma non credo che questa percentuale sia imputabile alla mia (nostra) inesperienza. Arrivare a oltre 10.000 strade approvate vuole dire molte ore/giorni di lavoro e molta esperienza acquisita.

Inoltre posso confermarti con grande certezza che il sistema rigetta alcune strade perfettamente legittime e corrette.

Concludo con il mio slogan personale in cui credo fermamente:

E, per chi è riuscito a leggere fino a qui, aggiungo un mio pensiero personale a quanto Morten ha scritto in un commento:

I feel that I somehow open myself up to a lot of criticism.

Da questo suo messaggio si capisce la voglia di crescere e di migliorarsi nonostante tutti i problemi che incontriamo e che, spesso, non capiamo. La condivisione di dubbi, perplessità o anche frustrazioni può solo aiutare se stessi e apportare vantaggi a tutti. Quindi non credo ci sia alcuno spazio a critiche.

Grazie di tutto

Take care

Marco

[PrasadVR]

@MarcoDavoli Thank you for sharing your experience. It’s very helpful.

I’m curious to know, Point 1. have you observed that approvals take longer when the existing white road shape corrected before adding a new road?

Point.2 You mentioned that organizing emails into folders in Gmail is an excellent solution. Could you please elaborate.

[MortenCopenhagen]

Thanks 1001 times for your wonderful reply, @MarcoDavoli .

It is filled with exactly the kind of knowledge sharing that I was hoping for. This way of learning from each other is perfect.

Thanks again

Morten

[MortenCopenhagen]

Hi @PrasadVR

Please see my first post in this series to learn about using filters and labels in Gmail.

Cheers

Morten

[PrasadVR]

@MortenCopenhagen Thank you for your kind clarification. I saw your post-1 and followed. I mistakenly thought he was referring to something else.

[MarcoDavoli]

Ciao @PrasadVR

among the different kind of road edit or modification sometimes happen that I can not resist and I do some “reshape” to an existing roads.

Unfortunately I never checked if this takes longer time to be approved. Also in the case where, after the re-shape, I also add a new road to it.

Have no tools to evaluate if takes longer time.

Take care

ready for the new TRAC Campaign?

ciao

Marco

[PrasadVR]

Ciao @MarcoDavoli , Thank you for your kind clarification. I felt, approval was taking long time when I changed the shape of the existing white road before adding a new road. I am excited for the meet-up and the TRAC.

[ErmesT]

Concordo @MarcoDavoli , non c’è posto per le critiche, ma solo per considerazioni e suggerimenti costruttivi. Il post di @MortenCopenhagen è interessante per confermare, come ho già detto, alcune delle mie osservazioni. Appena Maps aggiorna i dati dovrei ritrovarmi con oltre 28,000 strade approvate, di cui oltre 12,000 negli ultimi 60 giorni, con una media di 200 al giorno, in picco di 450 e in minimo di 100.

Chiarisco di nuovo che il mio 98% non tiene conto delle strade respinte dopo 90 giorni, in quanto si tratta di strade “non valutate” e che di solito la seconda volta vengono approvate in poche ore. Sto lavorando su in post al riguardo.

A parte quello che ho detto e scritto in precedenza sugli elementi che hanno priorità nell’approvazione, credo che ci siano strategie più o meno efficaci. La mia consiste in:

• Investire del tempo nell’individuazione delle aree da mappare e l’inserimento in una lista in Maps. Questo permette di poter lavorare intensivamente in pochi luoghi per volta.
• Inserimento in Maps di aree vaste, operando in 3 steps: Strade principali, diramazioni principali, diramazioni secondarie.

All’inizio può essere complicato e poco redditizio, ma nel giro di un mese si viene ripagati. Edits numerosi in un’area sembrano attirare l’attenzione della IA anche in aree poco abitate ma con una vasta espansione in corso.

Io faccio normalmente edit di shapes, ma solo se aggiungo una strada collegata. L’ edit della sola sagoma della strada non viene approvato mai.

La cancellazione di una strada viene accettata solo in combinazione con una strada aggiunta. Ho provato e funziona.

Sto studiando la gestione degli edit duplicati, ma per ora non ho una risposta.