Data mining / predictive stats courses are similar conceptually to a first/second course in linear regression, but expressed very differently than in standard econometrics. Adcoms will be concerned about your ability to transition into metrics-speak and pass first-year prelims; most students will have encountered the terms before and use their prior definitions to infer what the probabilistic/matrix expressions refer to, but you'll have nowhere to start. You'll also struggle with grad macro initially without diff equations. Grad micro may or may not be an issue depending on how rigorous your undergrad intermediate course is. I'd say the institution for the math/econ courses are less of an issue; 4.0/4.0 in a state flagship econ program and then a grad stats program is a very strong signal of problem-solving ability and work ethic. Problem is that you'll have some material to catch up on all three core sequences if you enroll next year, and top/mid-tier programs may be wary of providing you first-year funding if there's a high chance of you dropping/failing out. Some departments in public institutions are particularly concerned with attrition due to internal regulations/scrutiny.
I think 1 extra year of coursework may help a lot with your chances; you can possibly take some courses in econometrics and game theory while getting paid for RA work (which you also need more of, but this is a secondary concern for your range). But in either case, no harm in applying; I'd just suggest that if you don't get admitted with funding, take an extra year and aim for the same range of programs rather than settling for lower-ranked programs or an unfunded offer.