That’s a higher count than many. I’m on 21 which has taken “some effort”. You have 33 so you can imagine how much effort Paul HB9DST has put in to get to 100!
2 Likes
I can!!! My realistic life aim is 50. I’ll be pleased if I can make that. I have the time to do many more…. but not the funds.
What I need is a mega road trip around the USA…. but that is unlikely to ever happen. I have no plans to visit the USA currently. Although Canada might happen in the next couple of years and there’s 8 to go at there. I’ve also got family in Argentina. But that will mean getting a reciprocal licence - unfortunately, where they live in western Chubut province, isn’t yet an Association. Anyway…. that would give me all the continents bar Antarctica (which I don’t think will be realistic to achieve)
Don’t suppose there is an “Honour Roll” for Associations activated that’s hidden anywhere?
Anyway…. I’ve now hijacked this thread!
1 Like
I am receiving reports that the “recommender” returns nothing for some callsigns.
This being just an experiment, I did not implement/require the SOTA user login, which would uniquely identify the user.
Hence the many possible variations of a callsign used during activations (like ON6ZQ, ON6ZQ/P, DL/ON6ZQ/P, etc.) and “creative” schemes like in the UK (M0ABC, MW0ABC, MM0ABC, …) make it difficult to always identify correctly the correct UserID from the callsign only.
One example is “M5OTA” (UK) @M5OTA and its variations being confused with “TM5SOTA” (France), causing incorrect results.
If your callsign does not return any results and you do have activations stored in the SOTA database, let me know and I’ll run the program with the proper UserID and send you the results.
If people actually keep using this program, I might add SOTA authentication to resolve this issue once and for all, but then one will not be able to see recommendations for other callsigns anymore…
2 Likes
I tried it and it worked for me - interesting.
Apologies if I am missing something here, but i would be interested to know if the algorithm takes into account anything else other than the factors you mentioned.
73’s
David
1 Like
No.
This is a very basic experiment with no other information available than the activations logged in the SOTA database.
Basically,
- it looks for all summits that you have already activated,
- then looks for the summits that were activated by others who have activated the same summits as you,
- then calculates a score for the ones you have not yet activated but were activated by those who have activated the same summits as you.
I hope this makes sense.
Of course, using an history of your searches, posts, likes, purchase habits, address, birthdate, health condition, estimated wealth and a few pictures would greatly help making better recommendations, but this would go beyond the idea of this exercise 
Off topic, but this might surprise you:
3 Likes
I was expecting that complete canditates would be considered in the ranking?
Or at least if the user has a high complete rate?
1 Like
Not sure how to “weight” that. Not everyone looks for “complete”.
But the information is definitely useful when planning, so I added it like
| OE5JFE might also like |
|---|
| Summit Reference | Summit Name | chased | Score | WX |
|---|---|---|---|---|
| OK/JC-028 | Vítkův Kámen | 2024-03-17 | 3.78 | weather forecast |
| DM/BM-290 | Höhenberg | 3.33 | weather forecast | |
| OE/OO-119 | Göblberg | 2020-09-05 | 3.22 | weather forecast |
| HA/ND-001 | Irott-kö | 2018-04-29 | 3.18 | weather forecast |
| OE/SB-257 | Zwölferhorn | 2019-09-14 | 3.15 | weather forecast |
5 Likes
OH MY GOD…… I’ve just tried two random photos. That is properly scary!
Completely understand it not dealing with all callsign variations and overlaps. I’d seen your note on it and before I even tried it I was thinking the “creative” uk scheme must be a pain, never mind all the other variations to parse for something you are just doing for fun / interest 
Thanks for DMing me my list. Appreciate being able to join in the fun, although I didn’t want to put you to any extra effort. Lots on the list I’ve already got my eye on, but it is intriguing to see the other suggestions!
Based on my profile picture on QRZ…
His clothing and accessories hint at a penchant for outdoor activities such as hiking and amateur radio. But he also may excessively engage in binge-watching tv, drinking and gambling.
This man seems to possess moderate self-esteem and introversion, but his honesty and emotional stability are questionable.
Just want to make some corrections. I don’t gamble. 
5 Likes
A middle-aged man with grey hair is positioned in what appears to be a broadcasting studio. In the background are Euro 96 posters, and in the foreground, he wears headphones and points directly at the viewer, a microphone nearby. The setting suggests he’s about to deliver a statement or is in the midst of a live commentary.
This man, likely Caucasian, is probably in the £50,000 - £75,000 income range. We guess his religious affiliation is Christian, and he seems to be heterosexual. He is likely associated with the Conservative party. His biases may include ageism and sexism, and his racial biases might lean towards prejudice and tribalism. He looks confident and assertive, dressed in a sweatshirt, with interests in sports commentary, public speaking, and reading news but also showing negative interests towards online gambling, conspiracy theories and substance abuse.
This person seems to have high self-esteem and be quite extroverted, so we can target them with products that enhance their status and cater to their need for social interaction, such as Rode PodMic USB Microphone, Sony WH-1000XM5 Headphones, Bet365 Online Gambling, ExpressVPN, Rolex Submariner Watch, Tailored Business Suits from Savile Row, Luxury Sedan from BMW, Premium Scotch Whisky from The Macallan.
2 Likes
I made some improvements. Give it another try
1 Like
I see! Thank you 
Wow. Pretty close!
==
The man is likely Caucasian/New Zealander, with an estimated income range of NZD 50,000 to NZD 80,000. He probably does not follow a traditional religion, perhaps spiritual, identifies as heterosexual, and may lean towards the Green Party politically. Potential biases include a strong environmental bias and in-group favoritism. He shows happiness and contentment, wears an orange and black long-sleeved shirt and beige pants. His interests could be bird conservation, nature, and environmentalism, with potentially negative activities like hunting or fishing.
The man seems to possess strong self-esteem and is moderately extroverted. Therefore, we can target him with niche products like high-end bird watching gear and common items like outdoor clothing. Specific examples include the Swarovski Optik binoculars, Patagonia outdoor jackets, REI camping equipment, and Shimano fishing gear, alongside more general items such as Kathmandu hiking boots, donations to the Royal Forest and Bird Protection Society of New Zealand, subscriptions to National Geographic magazine, and eco-tourism packages from Intrepid Travel.
==
Mind you - the only photo of myself I could find of myself to test it with was one in which I’m holding a kiwi chick!
2 Likes
This was fun and identified a number of never-activated summits for me that are mostly do-able.
It reminds me of an online tool by someone who I’ve forgotten that allowed you to enter your call and another station’s call and it would tell you what they had activated that you had not.
This allowed you to mimic their planning. Or, to ‘keep up with the Joneses’, or do be able to say… ‘I’ll have what (s)he’s having’ to echo a famous movie scene.
One of the ways we all think about summits, and make decisions about how much effort to expend to get to a summit, is by the points. So I think differently about a 10-pointer verses a 1 or 2-pointer. Could you add a Points column to the results table?
There are other queries which could be bundled onto an Activator’s Tools page on SotaWatch. Another useful tool would be to show what I ‘normally’ activate in previous years that I have not activated yet this year. It’s not difficult to work out but a simple query tool would make that easy.
Well done, it’s fund to see creative uses of the SotaWatch API and the database we have personally and collectively logged now is a real asset.
2 Likes
Done.
2 Likes
Fun and interesting, it did indeed predict my #1 next target.
Thanks for adding summit points.
Here’s another observation, In my top 10 recommended summits, the ones highlighted in yellow are a considerable way away from my home QTH. For example, I would like to activate Mt Blue Rag but it is a 5 hr 30min drive away. Your algorithm is rightly picking it out because I’ve activated other peaks in the Victorian Alps around Mt Hotham and Falls Creek but not the Blue Rags. I expect to get up there one day but it requires a 2 or 3 night regional visit, which I have managed to do only half a dozen times in my 10 years of SOTA.
Whereas Mt Tarrengower is a 2 hr drive north-west, and Hickey is 1hr 40min – these two are eminently available and I should stop typing on a computer and go out and activate them!
So the algorithm does not take into account the distance, which impacts the feasibility of the activation. If it did, those highlighted summits would appear lower down in my list. Accessibility is another thing, but distance might be a simple approximation, recognising that if a reasonable number of others have activated a summit, then I should be able to as well.
Just a suggestion. Thanks for a fun piece of SOTA tooling!
2 Likes
That is a very good suggestion. It would help to indentify summits that are not only similar to past activity but also realistically reachable.
That said, the challenge is in defining “distance from where?”.
For example we could estimate an ‘activation centre’ from an activator’s past summits, or let users define their own reference (home grid square, ‘within X km from’, …)
Each has trade-offs between accuracy, simplicity and privacy.
Another consideration is balance: if distance dominates, the system turns into a “nearest summit” finder rather than a recommender of interesting, similar summits.
The best outcome probably would come from combining both: keeping the current “similarity” but applying a gentle distance weighting so “nearby” summits are prioritised without excluding distant but relevant ones.
Until we find a reliable and up to date source of data to do so, I suggest you sort the recommender results by clicking on the “Summit Reference” column header. This will already group the suggestions by SOTA Association/Region.
3 Likes
Distance from a weighted polygon enclosing past activations?
1 Like