Blog ranking software

We have developed an algorithm to rank blogs based on 9 different factors. Looking to automate the process of ranking the blogs instead of having to handle the ranking manually. Each factor will either happen by calling an external API (like PageRank) or by doing simple calculation. Historical data is saved as new calculation takes place, allowing analysis over time as the data set grows. Please indicated what technologies you will use. Finally, the pages will have to exist within Joomla. Additional requirements below. ## Deliverables Use Cases: 1. A user visits the site's main page and select a particular category, then a blog ranking will appear on the main page for each type of metric in the database. Each ranking will be limited to the top N (where N is a configurable value assumed to equal 25 during development) blogs. Any error in execution will result in a generic error message being displayed indicating technical difficulties, and the error itself will be logged. For example, the user comes to the system, and selects to view the top marketing blogs. The user is presented with a list of top 25 blogs in marketing. Please note that the number 25 is configurable. I should be able to change that number to 50 or 100 (via config file) and the system will display the top 100 blogs, etc. Look at a sample ranking for top SEO blogs here: [login to view URL] 2. blogs in each category are ranked based on the following: 1. RSS: Call to feedburner API to retrieve the number of rss subscribers. 2. PageRank: 3. Views / Month: Call to compete API to retrieve the views per month. framework coded, but actual call not yet. Need to get parameters from compete's website 4. Views / Visit:Call to compete API to retrieve the views per month. framework coded, but actual call not yet. Need to get parameters from compete's website 5. # of pages: Call to Yahoo's API. 6. # of inbound links: Call to Yahoo's API. 7. Links / Page: calculated based on two other factors. 8. Views / Page: calculated based on two other factors. 9. "Ultimate" metric calculated by applying weights to metric scores **1-6** 3. After the user views the top blogs in a certain category, he can add a new blog for ranking by submitting a URL and any other required information (such as blog url, feedburner url, categories for the blog). The user will also add his email so will be notified of any future changes in ranking. * If the URL is resolvable and can be successfully evaluated by all metrics, the Blog record is saved to the databased. The user will be able to see it's scores, and it will be included in future recalculations * If any error occurs, the user is informed and offered a chance to reenter the URL For example, the user views the top marketing blogs. The user is presented with a list of top 25 blogs in marketing. Then the user decides to add a new blog. The software will attempt to rank the blog in the categories based on the same above factors. If ranking is successful, the top 25 ranking might change or the user might be told that is ranked as number xx based on each of the factors. 4. A blog can belong to "X" number of categories where X is configurable. When the user submits a blog to be ranked, he can select up to a number of categories. For examlpe, maybe 5 different categories. 5. If a user is viewing a list of blogs in cetain category, then he can click on one particular blog and he will be presented with the data captured for that blog by itself. For example, the user will view for one blog: 1. RSS: 2. PageRank: 3. Views / Month 4. other factors 6. Once a week, a cron job will run to rank blogs in all different categories. I don't want calculation for top blogs to happen real time. Rather, when a user wants to view top blogs in a certain category, he will be presented with the top blogs based on our last caculation. 7. Maintain a history for each blog indicating the rank for each factor on weekly basis. As the cron runs ever week to obtain new values for each blog, we need to insert the previous records to a history table. So, if a blog has 1000 RSS subscribers in week one, then in week two, we run the cron, we find he has 1200 subscribers. We need to store the 1000 subscribers in the history table. 8. Allow a blog owner to take some Java script code and put it on their site to display their ranking in a certain category. 9. Errors will be logged to standard out. **Answers to additional questions I received:** >>A user visits the site's main page and select a particular category, then a blog ranking will appear on the main page for each type of metric in the database. Can you please explain more about this? 1a. What site's main page contains BEFORE user will select a particular category? I need to be more accurate. The will visit the application main page. Let's say [login to view URL] He will see a drop down menu that will contain a list of the different categories. He can select a category, press submit and he will be taken to the page that shows the ranking for that category. 1b. Each type of metric - what type of metrics do you have? Do you want many tables to appear on the main page like "TOP 25 BY METRIC A", "TOP 25 BY METRIC B" etc. At this point, we will have one pages for results. One , after the user selects a category, he will be taken to ranking page: top blogs in category x. The page will have top blogs in category x by rss top blogs in category x by links etc based on the factors mentioned in the requirements. From that page, a user can decide to add a new blog to the list. 2. Here is your example about top SEO blogs: >>Look at a sample ranking for top SEO blogs here [login to view URL] At this URL I can see some text before the list - must I put any text like this to the "TOP X..." page or not? you don't have to worry about putting the text. 3. In ranking process you mention about "Call to compete API" - what is this? Can you describe more about this? [login to view URL] provides a way to get an estimate of monthly views and pages per visit for a domain. Both of these are factors in the ranking. 4. Questions about ability for users to add their blogs: 4a. Whether user must be registered first, and then submit one or more blogs using his account? No. Anyone should be able to add a blog. 4b. Must user confirm his email first (like usually it happens in registration process), and only then be able to submit blogs? No. Anyone should be able to add a blog. 4c. You wrote: >>...by submitting a URL and any other required information... Can you give more details: what exactly fields do you need them to submit; what fields are obligatory and what are not. We will need the url for the blog, feedburner url, user email and a checkbox to say "add me to list" 5. User membership questions: 5a. Do you want users to be able to log in to their account and to view the list of submitted blogs and the status of blogs? No. 5b. If answer to 5a is YES, then... what information do you want to show to the users in their accounts? Maybe you want to show them the actual scores, maybe also the scores history? Maybe with graphs or maybe tables are enough? 6. Questions about calculations. 6a. I propose you to organize it so the calculations will be not performed every week, but will be distributed on all days of the week. Eg, if you have 10000 blogs in your database, 1429 blogs will be calculated every day using cron. Do you like this idea? There are two aspects to the calculation. 1. One is sending requests to external API to retrieve data(like RSS membership, google rank, etc). That should happen once a week per blog. Of course the actual cron that runs that will have to run continuesly. The interval of how frequently we send a request should be configurable. So initially, it might 7 days but if the list of blogs grows too large, we might want to run the cron once a month. Also, when the cron collects new data for a blog, it needs to insert its current data to history table. 2. category calculation based on the collected data from cron 1 above. We should run that once a week per category after all blogs in that category have their data collected. 6b. As about history table - I have different concept of how it will work inside using database, but you will get the same result. We can discuss it later if you accept my bid. I am assuming that you will have a blog table where you keep current score for each of the factors. History table will mirror current table but adds date ranges as well. I do have the initial scheme created for this and will be glad to share it. 7. Besides blogs ranking and showing top X blogs functionality, what other features do you need on the site? Eg places for advertising, any admin panel to configure it, or config file is enough, news, articles etc - just enumerate all features that you need besides ranking functionality. This all must be described in specification and will vary the price of the project. None. At this initial phase, I would like get a simple functionality out. At a later release, we might add premium functionality. 8. Do you know that there is no API to get Google Page Rank value? But I have the algorithm on how to query and decode this value.... or they have added new API to get Google PR? It seems no. Yep, I know there is no api but you can use the hash method to do the calculation. What is important is to remember that we will be adding more functionality to the app later on, so I want whatever code developed to be flexible enough for future addition. Also, I will need to get a quick tech design doc to outline modules in the system and how they will be implemented. I come from a Java background but I think I can figure PHP.